1 2 3 The Project Gutenberg Etext of LOC WORKSHOP ON ELECTRONIC TEXTS 4 5 6 7 8 WORKSHOP ON ELECTRONIC TEXTS 9 10 PROCEEDINGS 11 12 13 14 Edited by James Daly 15 16 17 18 19 20 21 22 9-10 June 1992 23 24 25 Library of Congress 26 Washington, D.C. 27 28 29 30 Supported by a Grant from the David and Lucile Packard Foundation 31 32 33 *** *** *** ****** *** *** *** 34 35 36 TABLE OF CONTENTS 37 38 39 Acknowledgements 40 41 Introduction 42 43 Proceedings 44 Welcome 45 Prosser Gifford and Carl Fleischhauer 46 47 Session I. Content in a New Form: Who Will Use It and What Will They Do? 48 James Daly (Moderator) 49 Avra Michelson, Overview 50 Susan H. Veccia, User Evaluation 51 Joanne Freeman, Beyond the Scholar 52 Discussion 53 54 Session II. Show and Tell 55 Jacqueline Hess (Moderator) 56 Elli Mylonas, Perseus Project 57 Discussion 58 Eric M. Calaluca, Patrologia Latina Database 59 Carl Fleischhauer and Ricky Erway, American Memory 60 Discussion 61 Dorothy Twohig, The Papers of George Washington 62 Discussion 63 Maria L. Lebron, The Online Journal of Current Clinical Trials 64 Discussion 65 Lynne K. Personius, Cornell mathematics books 66 Discussion 67 68 Session III. Distribution, Networks, and Networking: 69 Options for Dissemination 70 Robert G. Zich (Moderator) 71 Clifford A. Lynch 72 Discussion 73 Howard Besser 74 Discussion 75 Ronald L. Larsen 76 Edwin B. Brownrigg 77 Discussion 78 79 Session IV. Image Capture, Text Capture, Overview of Text and 80 Image Storage Formats 81 William L. Hooton (Moderator) 82 A) Principal Methods for Image Capture of Text: 83 direct scanning, use of microform 84 Anne R. Kenney 85 Pamela Q.J. Andre 86 Judith A. Zidar 87 Donald J. Waters 88 Discussion 89 B) Special Problems: bound volumes, conservation, 90 reproducing printed halftones 91 George Thoma 92 Carl Fleischhauer 93 Discussion 94 C) Image Standards and Implications for Preservation 95 Jean Baronas 96 Patricia Battin 97 Discussion 98 D) Text Conversion: OCR vs. rekeying, standards of accuracy 99 and use of imperfect texts, service bureaus 100 Michael Lesk 101 Ricky Erway 102 Judith A. Zidar 103 Discussion 104 105 Session V. Approaches to Preparing Electronic Texts 106 Susan Hockey (Moderator) 107 Stuart Weibel 108 Discussion 109 C.M. Sperberg-McQueen 110 Discussion 111 Eric M. Calaluca 112 Discussion 113 114 Session VI. Copyright Issues 115 Marybeth Peters 116 117 Session VII. Conclusion 118 Prosser Gifford (Moderator) 119 General discussion 120 121 Appendix I: Program 122 123 Appendix II: Abstracts 124 125 Appendix III: Directory of Participants 126 127 128 *** *** *** ****** *** *** *** 129 130 131 Acknowledgements 132 133 I would like to thank Carl Fleischhauer and Prosser Gifford for the 134 opportunity to learn about areas of human activity unknown to me a scant 135 ten months ago, and the David and Lucile Packard Foundation for 136 supporting that opportunity. The help given by others is acknowledged on 137 a separate page. 138 139 19 October 1992 140 141 142 *** *** *** ****** *** *** *** 143 144 145 INTRODUCTION 146 147 The Workshop on Electronic Texts (1) drew together representatives of 148 various projects and interest groups to compare ideas, beliefs, 149 experiences, and, in particular, methods of placing and presenting 150 historical textual materials in computerized form. Most attendees gained 151 much in insight and outlook from the event. But the assembly did not 152 form a new nation, or, to put it another way, the diversity of projects 153 and interests was too great to draw the representatives into a cohesive, 154 action-oriented body.(2) 155 156 Everyone attending the Workshop shared an interest in preserving and 157 providing access to historical texts. But within this broad field the 158 attendees represented a variety of formal, informal, figurative, and 159 literal groups, with many individuals belonging to more than one. These 160 groups may be defined roughly according to the following topics or 161 activities: 162 163 * Imaging 164 * Searchable coded texts 165 * National and international computer networks 166 * CD-ROM production and dissemination 167 * Methods and technology for converting older paper materials into 168 electronic form 169 * Study of the use of digital materials by scholars and others 170 171 This summary is arranged thematically and does not follow the actual 172 sequence of presentations. 173 174 NOTES: 175 (1) In this document, the phrase electronic text is used to mean 176 any computerized reproduction or version of a document, book, 177 article, or manuscript (including images), and not merely a machine- 178 readable or machine-searchable text. 179 180 (2) The Workshop was held at the Library of Congress on 9-10 June 181 1992, with funding from the David and Lucile Packard Foundation. 182 The document that follows represents a summary of the presentations 183 made at the Workshop and was compiled by James DALY. This 184 introduction was written by DALY and Carl FLEISCHHAUER. 185 186 187 PRESERVATION AND IMAGING 188 189 Preservation, as that term is used by archivists,(3) was most explicitly 190 discussed in the context of imaging. Anne KENNEY and Lynne PERSONIUS 191 explained how the concept of a faithful copy and the user-friendliness of 192 the traditional book have guided their project at Cornell University.(4) 193 Although interested in computerized dissemination, participants in the 194 Cornell project are creating digital image sets of older books in the 195 public domain as a source for a fresh paper facsimile or, in a future 196 phase, microfilm. The books returned to the library shelves are 197 high-quality and useful replacements on acid-free paper that should last 198 a long time. To date, the Cornell project has placed little or no 199 emphasis on creating searchable texts; one would not be surprised to find 200 that the project participants view such texts as new editions, and thus 201 not as faithful reproductions. 202 203 In her talk on preservation, Patricia BATTIN struck an ecumenical and 204 flexible note as she endorsed the creation and dissemination of a variety 205 of types of digital copies. Do not be too narrow in defining what counts 206 as a preservation element, BATTIN counseled; for the present, at least, 207 digital copies made with preservation in mind cannot be as narrowly 208 standardized as, say, microfilm copies with the same objective. Setting 209 standards precipitously can inhibit creativity, but delay can result in 210 chaos, she advised. 211 212 In part, BATTIN's position reflected the unsettled nature of image-format 213 standards, and attendees could hear echoes of this unsettledness in the 214 comments of various speakers. For example, Jean BARONAS reviewed the 215 status of several formal standards moving through committees of experts; 216 and Clifford LYNCH encouraged the use of a new guideline for transmitting 217 document images on Internet. Testimony from participants in the National 218 Agricultural Library's (NAL) Text Digitization Program and LC's American 219 Memory project highlighted some of the challenges to the actual creation 220 or interchange of images, including difficulties in converting 221 preservation microfilm to digital form. Donald WATERS reported on the 222 progress of a master plan for a project at Yale University to convert 223 books on microfilm to digital image sets, Project Open Book (POB). 224 225 The Workshop offered rather less of an imaging practicum than planned, 226 but "how-to" hints emerge at various points, for example, throughout 227 KENNEY's presentation and in the discussion of arcana such as 228 thresholding and dithering offered by George THOMA and FLEISCHHAUER. 229 230 NOTES: 231 (3) Although there is a sense in which any reproductions of 232 historical materials preserve the human record, specialists in the 233 field have developed particular guidelines for the creation of 234 acceptable preservation copies. 235 236 (4) Titles and affiliations of presenters are given at the 237 beginning of their respective talks and in the Directory of 238 Participants (Appendix III). 239 240 241 THE MACHINE-READABLE TEXT: MARKUP AND USE 242 243 The sections of the Workshop that dealt with machine-readable text tended 244 to be more concerned with access and use than with preservation, at least 245 in the narrow technical sense. Michael SPERBERG-McQUEEN made a forceful 246 presentation on the Text Encoding Initiative's (TEI) implementation of 247 the Standard Generalized Markup Language (SGML). His ideas were echoed 248 by Susan HOCKEY, Elli MYLONAS, and Stuart WEIBEL. While the 249 presentations made by the TEI advocates contained no practicum, their 250 discussion focused on the value of the finished product, what the 251 European Community calls reusability, but what may also be termed 252 durability. They argued that marking up--that is, coding--a text in a 253 well-conceived way will permit it to be moved from one computer 254 environment to another, as well as to be used by various users. Two 255 kinds of markup were distinguished: 1) procedural markup, which 256 describes the features of a text (e.g., dots on a page), and 2) 257 descriptive markup, which describes the structure or elements of a 258 document (e.g., chapters, paragraphs, and front matter). 259 260 The TEI proponents emphasized the importance of texts to scholarship. 261 They explained how heavily coded (and thus analyzed and annotated) texts 262 can underlie research, play a role in scholarly communication, and 263 facilitate classroom teaching. SPERBERG-McQUEEN reminded listeners that 264 a written or printed item (e.g., a particular edition of a book) is 265 merely a representation of the abstraction we call a text. To concern 266 ourselves with faithfully reproducing a printed instance of the text, 267 SPERBERG-McQUEEN argued, is to concern ourselves with the representation 268 of a representation ("images as simulacra for the text"). The TEI proponents' 269 interest in images tends to focus on corollary materials for use in teaching, 270 for example, photographs of the Acropolis to accompany a Greek text. 271 272 By the end of the Workshop, SPERBERG-McQUEEN confessed to having been 273 converted to a limited extent to the view that electronic images 274 constitute a promising alternative to microfilming; indeed, an 275 alternative probably superior to microfilming. But he was not convinced 276 that electronic images constitute a serious attempt to represent text in 277 electronic form. HOCKEY and MYLONAS also conceded that their experience 278 at the Pierce Symposium the previous week at Georgetown University and 279 the present conference at the Library of Congress had compelled them to 280 reevaluate their perspective on the usefulness of text as images. 281 Attendees could see that the text and image advocates were in 282 constructive tension, so to say. 283 284 Three nonTEI presentations described approaches to preparing 285 machine-readable text that are less rigorous and thus less expensive. In 286 the case of the Papers of George Washington, Dorothy TWOHIG explained 287 that the digital version will provide a not-quite-perfect rendering of 288 the transcribed text--some 135,000 documents, available for research 289 during the decades while the perfect or print version is completed. 290 Members of the American Memory team and the staff of NAL's Text 291 Digitization Program (see below) also outlined a middle ground concerning 292 searchable texts. In the case of American Memory, contractors produce 293 texts with about 99-percent accuracy that serve as "browse" or 294 "reference" versions of written or printed originals. End users who need 295 faithful copies or perfect renditions must refer to accompanying sets of 296 digital facsimile images or consult copies of the originals in a nearby 297 library or archive. American Memory staff argued that the high cost of 298 producing 100-percent accurate copies would prevent LC from offering 299 access to large parts of its collections. 300 301 302 THE MACHINE-READABLE TEXT: METHODS OF CONVERSION 303 304 Although the Workshop did not include a systematic examination of the 305 methods for converting texts from paper (or from facsimile images) into 306 machine-readable form, nevertheless, various speakers touched upon this 307 matter. For example, WEIBEL reported that OCLC has experimented with a 308 merging of multiple optical character recognition systems that will 309 reduce errors from an unacceptable rate of 5 characters out of every 310 l,000 to an unacceptable rate of 2 characters out of every l,000. 311 312 Pamela ANDRE presented an overview of NAL's Text Digitization Program and 313 Judith ZIDAR discussed the technical details. ZIDAR explained how NAL 314 purchased hardware and software capable of performing optical character 315 recognition (OCR) and text conversion and used its own staff to convert 316 texts. The process, ZIDAR said, required extensive editing and project 317 staff found themselves considering alternatives, including rekeying 318 and/or creating abstracts or summaries of texts. NAL reckoned costs at 319 $7 per page. By way of contrast, Ricky ERWAY explained that American 320 Memory had decided from the start to contract out conversion to external 321 service bureaus. The criteria used to select these contractors were cost 322 and quality of results, as opposed to methods of conversion. ERWAY noted 323 that historical documents or books often do not lend themselves to OCR. 324 Bound materials represent a special problem. In her experience, quality 325 control--inspecting incoming materials, counting errors in samples--posed 326 the most time-consuming aspect of contracting out conversion. ERWAY 327 reckoned American Memory's costs at $4 per page, but cautioned that fewer 328 cost-elements had been included than in NAL's figure. 329 330 331 OPTIONS FOR DISSEMINATION 332 333 The topic of dissemination proper emerged at various points during the 334 Workshop. At the session devoted to national and international computer 335 networks, LYNCH, Howard BESSER, Ronald LARSEN, and Edwin BROWNRIGG 336 highlighted the virtues of Internet today and of the network that will 337 evolve from Internet. Listeners could discern in these narratives a 338 vision of an information democracy in which millions of citizens freely 339 find and use what they need. LYNCH noted that a lack of standards 340 inhibits disseminating multimedia on the network, a topic also discussed 341 by BESSER. LARSEN addressed the issues of network scalability and 342 modularity and commented upon the difficulty of anticipating the effects 343 of growth in orders of magnitude. BROWNRIGG talked about the ability of 344 packet radio to provide certain links in a network without the need for 345 wiring. However, the presenters also called attention to the 346 shortcomings and incongruities of present-day computer networks. For 347 example: 1) Network use is growing dramatically, but much network 348 traffic consists of personal communication (E-mail). 2) Large bodies of 349 information are available, but a user's ability to search across their 350 entirety is limited. 3) There are significant resources for science and 351 technology, but few network sources provide content in the humanities. 352 4) Machine-readable texts are commonplace, but the capability of the 353 system to deal with images (let alone other media formats) lags behind. 354 A glimpse of a multimedia future for networks, however, was provided by 355 Maria LEBRON in her overview of the Online Journal of Current Clinical 356 Trials (OJCCT), and the process of scholarly publishing on-line. 357 358 The contrasting form of the CD-ROM disk was never systematically 359 analyzed, but attendees could glean an impression from several of the 360 show-and-tell presentations. The Perseus and American Memory examples 361 demonstrated recently published disks, while the descriptions of the 362 IBYCUS version of the Papers of George Washington and Chadwyck-Healey's 363 Patrologia Latina Database (PLD) told of disks to come. According to 364 Eric CALALUCA, PLD's principal focus has been on converting Jacques-Paul 365 Migne's definitive collection of Latin texts to machine-readable form. 366 Although everyone could share the network advocates' enthusiasm for an 367 on-line future, the possibility of rolling up one's sleeves for a session 368 with a CD-ROM containing both textual materials and a powerful retrieval 369 engine made the disk seem an appealing vessel indeed. The overall 370 discussion suggested that the transition from CD-ROM to on-line networked 371 access may prove far slower and more difficult than has been anticipated. 372 373 374 WHO ARE THE USERS AND WHAT DO THEY DO? 375 376 Although concerned with the technicalities of production, the Workshop 377 never lost sight of the purposes and uses of electronic versions of 378 textual materials. As noted above, those interested in imaging discussed 379 the problematical matter of digital preservation, while the TEI proponents 380 described how machine-readable texts can be used in research. This latter 381 topic received thorough treatment in the paper read by Avra MICHELSON. 382 She placed the phenomenon of electronic texts within the context of 383 broader trends in information technology and scholarly communication. 384 385 Among other things, MICHELSON described on-line conferences that 386 represent a vigorous and important intellectual forum for certain 387 disciplines. Internet now carries more than 700 conferences, with about 388 80 percent of these devoted to topics in the social sciences and the 389 humanities. Other scholars use on-line networks for "distance learning." 390 Meanwhile, there has been a tremendous growth in end-user computing; 391 professors today are less likely than their predecessors to ask the 392 campus computer center to process their data. Electronic texts are one 393 key to these sophisticated applications, MICHELSON reported, and more and 394 more scholars in the humanities now work in an on-line environment. 395 Toward the end of the Workshop, Michael LESK presented a corollary to 396 MICHELSON's talk, reporting the results of an experiment that compared 397 the work of one group of chemistry students using traditional printed 398 texts and two groups using electronic sources. The experiment 399 demonstrated that in the event one does not know what to read, one needs 400 the electronic systems; the electronic systems hold no advantage at the 401 moment if one knows what to read, but neither do they impose a penalty. 402 403 DALY provided an anecdotal account of the revolutionizing impact of the 404 new technology on his previous methods of research in the field of classics. 405 His account, by extrapolation, served to illustrate in part the arguments 406 made by MICHELSON concerning the positive effects of the sudden and radical 407 transformation being wrought in the ways scholars work. 408 409 Susan VECCIA and Joanne FREEMAN delineated the use of electronic 410 materials outside the university. The most interesting aspect of their 411 use, FREEMAN said, could be seen as a paradox: teachers in elementary 412 and secondary schools requested access to primary source materials but, 413 at the same time, found that "primariness" itself made these materials 414 difficult for their students to use. 415 416 417 OTHER TOPICS 418 419 Marybeth PETERS reviewed copyright law in the United States and offered 420 advice during a lively discussion of this subject. But uncertainty 421 remains concerning the price of copyright in a digital medium, because a 422 solution remains to be worked out concerning management and synthesis of 423 copyrighted and out-of-copyright pieces of a database. 424 425 As moderator of the final session of the Workshop, Prosser GIFFORD directed 426 discussion to future courses of action and the potential role of LC in 427 advancing them. Among the recommendations that emerged were the following: 428 429 * Workshop participants should 1) begin to think about working 430 with image material, but structure and digitize it in such a 431 way that at a later stage it can be interpreted into text, and 432 2) find a common way to build text and images together so that 433 they can be used jointly at some stage in the future, with 434 appropriate network support, because that is how users will want 435 to access these materials. The Library might encourage attempts 436 to bring together people who are working on texts and images. 437 438 * A network version of American Memory should be developed or 439 consideration should be given to making the data in it 440 available to people interested in doing network multimedia. 441 Given the current dearth of digital data that is appealing and 442 unencumbered by extremely complex rights problems, developing a 443 network version of American Memory could do much to help make 444 network multimedia a reality. 445 446 * Concerning the thorny issue of electronic deposit, LC should 447 initiate a catalytic process in terms of distributed 448 responsibility, that is, bring together the distributed 449 organizations and set up a study group to look at all the 450 issues related to electronic deposit and see where we as a 451 nation should move. For example, LC might attempt to persuade 452 one major library in each state to deal with its state 453 equivalent publisher, which might produce a cooperative project 454 that would be equitably distributed around the country, and one 455 in which LC would be dealing with a minimal number of publishers 456 and minimal copyright problems. LC must also deal with the 457 concept of on-line publishing, determining, among other things, 458 how serials such as OJCCT might be deposited for copyright. 459 460 * Since a number of projects are planning to carry out 461 preservation by creating digital images that will end up in 462 on-line or near-line storage at some institution, LC might play 463 a helpful role, at least in the near term, by accelerating how 464 to catalog that information into the Research Library Information 465 Network (RLIN) and then into OCLC, so that it would be accessible. 466 This would reduce the possibility of multiple institutions digitizing 467 the same work. 468 469 470 CONCLUSION 471 472 The Workshop was valuable because it brought together partisans from 473 various groups and provided an occasion to compare goals and methods. 474 The more committed partisans frequently communicate with others in their 475 groups, but less often across group boundaries. The Workshop was also 476 valuable to attendees--including those involved with American Memory--who 477 came less committed to particular approaches or concepts. These 478 attendees learned a great deal, and plan to select and employ elements of 479 imaging, text-coding, and networked distribution that suit their 480 respective projects and purposes. 481 482 Still, reality rears its ugly head: no breakthrough has been achieved. 483 On the imaging side, one confronts a proliferation of competing 484 data-interchange standards and a lack of consensus on the role of digital 485 facsimiles in preservation. In the realm of machine-readable texts, one 486 encounters a reasonably mature standard but methodological difficulties 487 and high costs. These latter problems, of course, represent a special 488 impediment to the desire, as it is sometimes expressed in the popular 489 press, "to put the [contents of the] Library of Congress on line." In 490 the words of one participant, there was "no solution to the economic 491 problems--the projects that are out there are surviving, but it is going 492 to be a lot of work to transform the information industry, and so far the 493 investment to do that is not forthcoming" (LESK, per litteras). 494 495 496 *** *** *** ****** *** *** *** 497 498 499 PROCEEDINGS 500 501 502 WELCOME 503 504 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 505 GIFFORD * Origin of Workshop in current Librarian's desire to make LC's 506 collections more widely available * Desiderata arising from the prospect 507 of greater interconnectedness * 508 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 509 510 After welcoming participants on behalf of the Library of Congress, 511 American Memory (AM), and the National Demonstration Lab, Prosser 512 GIFFORD, director for scholarly programs, Library of Congress, located 513 the origin of the Workshop on Electronic Texts in a conversation he had 514 had considerably more than a year ago with Carl FLEISCHHAUER concerning 515 some of the issues faced by AM. On the assumption that numerous other 516 people were asking the same questions, the decision was made to bring 517 together as many of these people as possible to ask the same questions 518 together. In a deeper sense, GIFFORD said, the origin of the Workshop 519 lay in the desire of the current Librarian of Congress, James H. 520 Billington, to make the collections of the Library, especially those 521 offering unique or unusual testimony on aspects of the American 522 experience, available to a much wider circle of users than those few 523 people who can come to Washington to use them. This meant that the 524 emphasis of AM, from the outset, has been on archival collections of the 525 basic material, and on making these collections themselves available, 526 rather than selected or heavily edited products. 527 528 From AM's emphasis followed the questions with which the Workshop began: 529 who will use these materials, and in what form will they wish to use 530 them. But an even larger issue deserving mention, in GIFFORD's view, was 531 the phenomenal growth in Internet connectivity. He expressed the hope 532 that the prospect of greater interconnectedness than ever before would 533 lead to: 1) much more cooperative and mutually supportive endeavors; 2) 534 development of systems of shared and distributed responsibilities to 535 avoid duplication and to ensure accuracy and preservation of unique 536 materials; and 3) agreement on the necessary standards and development of 537 the appropriate directories and indices to make navigation 538 straightforward among the varied resources that are, and increasingly 539 will be, available. In this connection, GIFFORD requested that 540 participants reflect from the outset upon the sorts of outcomes they 541 thought the Workshop might have. Did those present constitute a group 542 with sufficient common interests to propose a next step or next steps, 543 and if so, what might those be? They would return to these questions the 544 following afternoon. 545 546 ****** 547 548 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 549 FLEISCHHAUER * Core of Workshop concerns preparation and production of 550 materials * Special challenge in conversion of textual materials * 551 Quality versus quantity * Do the several groups represented share common 552 interests? * 553 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 554 555 Carl FLEISCHHAUER, coordinator, American Memory, Library of Congress, 556 emphasized that he would attempt to represent the people who perform some 557 of the work of converting or preparing materials and that the core of 558 the Workshop had to do with preparation and production. FLEISCHHAUER 559 then drew a distinction between the long term, when many things would be 560 available and connected in the ways that GIFFORD described, and the short 561 term, in which AM not only has wrestled with the issue of what is the 562 best course to pursue but also has faced a variety of technical 563 challenges. 564 565 FLEISCHHAUER remarked AM's endeavors to deal with a wide range of library 566 formats, such as motion picture collections, sound-recording collections, 567 and pictorial collections of various sorts, especially collections of 568 photographs. In the course of these efforts, AM kept coming back to 569 textual materials--manuscripts or rare printed matter, bound materials, 570 etc. Text posed the greatest conversion challenge of all. Thus, the 571 genesis of the Workshop, which reflects the problems faced by AM. These 572 problems include physical problems. For example, those in the library 573 and archive business deal with collections made up of fragile and rare 574 manuscript items, bound materials, especially the notoriously brittle 575 bound materials of the late nineteenth century. These are precious 576 cultural artifacts, however, as well as interesting sources of 577 information, and LC desires to retain and conserve them. AM needs to 578 handle things without damaging them. Guillotining a book to run it 579 through a sheet feeder must be avoided at all costs. 580 581 Beyond physical problems, issues pertaining to quality arose. For 582 example, the desire to provide users with a searchable text is affected 583 by the question of acceptable level of accuracy. One hundred percent 584 accuracy is tremendously expensive. On the other hand, the output of 585 optical character recognition (OCR) can be tremendously inaccurate. 586 Although AM has attempted to find a middle ground, uncertainty persists 587 as to whether or not it has discovered the right solution. 588 589 Questions of quality arose concerning images as well. FLEISCHHAUER 590 contrasted the extremely high level of quality of the digital images in 591 the Cornell Xerox Project with AM's efforts to provide a browse-quality 592 or access-quality image, as opposed to an archival or preservation image. 593 FLEISCHHAUER therefore welcomed the opportunity to compare notes. 594 595 FLEISCHHAUER observed in passing that conversations he had had about 596 networks have begun to signal that for various forms of media a 597 determination may be made that there is a browse-quality item, or a 598 distribution-and-access-quality item that may coexist in some systems 599 with a higher quality archival item that would be inconvenient to send 600 through the network because of its size. FLEISCHHAUER referred, of 601 course, to images more than to searchable text. 602 603 As AM considered those questions, several conceptual issues arose: ought 604 AM occasionally to reproduce materials entirely through an image set, at 605 other times, entirely through a text set, and in some cases, a mix? 606 There probably would be times when the historical authenticity of an 607 artifact would require that its image be used. An image might be 608 desirable as a recourse for users if one could not provide 100-percent 609 accurate text. Again, AM wondered, as a practical matter, if a 610 distinction could be drawn between rare printed matter that might exist 611 in multiple collections--that is, in ten or fifteen libraries. In such 612 cases, the need for perfect reproduction would be less than for unique 613 items. Implicit in his remarks, FLEISCHHAUER conceded, was the admission 614 that AM has been tilting strongly towards quantity and drawing back a 615 little from perfect quality. That is, it seemed to AM that society would 616 be better served if more things were distributed by LC--even if they were 617 not quite perfect--than if fewer things, perfectly represented, were 618 distributed. This was stated as a proposition to be tested, with 619 responses to be gathered from users. 620 621 In thinking about issues related to reproduction of materials and seeing 622 other people engaged in parallel activities, AM deemed it useful to 623 convene a conference. Hence, the Workshop. FLEISCHHAUER thereupon 624 surveyed the several groups represented: 1) the world of images (image 625 users and image makers); 2) the world of text and scholarship and, within 626 this group, those concerned with language--FLEISCHHAUER confessed to finding 627 delightful irony in the fact that some of the most advanced thinkers on 628 computerized texts are those dealing with ancient Greek and Roman materials; 629 3) the network world; and 4) the general world of library science, which 630 includes people interested in preservation and cataloging. 631 632 FLEISCHHAUER concluded his remarks with special thanks to the David and 633 Lucile Packard Foundation for its support of the meeting, the American 634 Memory group, the Office for Scholarly Programs, the National 635 Demonstration Lab, and the Office of Special Events. He expressed the 636 hope that David Woodley Packard might be able to attend, noting that 637 Packard's work and the work of the foundation had sponsored a number of 638 projects in the text area. 639 640 ****** 641 642 SESSION I. CONTENT IN A NEW FORM: WHO WILL USE IT AND WHAT WILL THEY DO? 643 644 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 645 DALY * Acknowledgements * A new Latin authors disk * Effects of the new 646 technology on previous methods of research * 647 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 648 649 Serving as moderator, James DALY acknowledged the generosity of all the 650 presenters for giving of their time, counsel, and patience in planning 651 the Workshop, as well as of members of the American Memory project and 652 other Library of Congress staff, and the David and Lucile Packard 653 Foundation and its executive director, Colburn S. Wilbur. 654 655 DALY then recounted his visit in March to the Center for Electronic Texts 656 in the Humanities (CETH) and the Department of Classics at Rutgers 657 University, where an old friend, Lowell Edmunds, introduced him to the 658 department's IBYCUS scholarly personal computer, and, in particular, the 659 new Latin CD-ROM, containing, among other things, almost all classical 660 Latin literary texts through A.D. 200. Packard Humanities Institute 661 (PHI), Los Altos, California, released this disk late in 1991, with a 662 nominal triennial licensing fee. 663 664 Playing with the disk for an hour or so at Rutgers brought home to DALY 665 at once the revolutionizing impact of the new technology on his previous 666 methods of research. Had this disk been available two or three years 667 earlier, DALY contended, when he was engaged in preparing a commentary on 668 Book 10 of Virgil's Aeneid for Cambridge University Press, he would not 669 have required a forty-eight-square-foot table on which to spread the 670 numerous, most frequently consulted items, including some ten or twelve 671 concordances to key Latin authors, an almost equal number of lexica to 672 authors who lacked concordances, and where either lexica or concordances 673 were lacking, numerous editions of authors antedating and postdating Virgil. 674 675 Nor, when checking each of the average six to seven words contained in 676 the Virgilian hexameter for its usage elsewhere in Virgil's works or 677 other Latin authors, would DALY have had to maintain the laborious 678 mechanical process of flipping through these concordances, lexica, and 679 editions each time. Nor would he have had to frequent as often the 680 Milton S. Eisenhower Library at the Johns Hopkins University to consult 681 the Thesaurus Linguae Latinae. Instead of devoting countless hours, or 682 the bulk of his research time, to gathering data concerning Virgil's use 683 of words, DALY--now freed by PHI's Latin authors disk from the 684 tyrannical, yet in some ways paradoxically happy scholarly drudgery-- 685 would have been able to devote that same bulk of time to analyzing and 686 interpreting Virgilian verbal usage. 687 688 Citing Theodore Brunner, Gregory Crane, Elli MYLONAS, and Avra MICHELSON, 689 DALY argued that this reversal in his style of work, made possible by the 690 new technology, would perhaps have resulted in better, more productive 691 research. Indeed, even in the course of his browsing the Latin authors 692 disk at Rutgers, its powerful search, retrieval, and highlighting 693 capabilities suggested to him several new avenues of research into 694 Virgil's use of sound effects. This anecdotal account, DALY maintained, 695 may serve to illustrate in part the sudden and radical transformation 696 being wrought in the ways scholars work. 697 698 ****** 699 700 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 701 MICHELSON * Elements related to scholarship and technology * Electronic 702 texts within the context of broader trends within information technology 703 and scholarly communication * Evaluation of the prospects for the use of 704 electronic texts * Relationship of electronic texts to processes of 705 scholarly communication in humanities research * New exchange formats 706 created by scholars * Projects initiated to increase scholarly access to 707 converted text * Trend toward making electronic resources available 708 through research and education networks * Changes taking place in 709 scholarly communication among humanities scholars * Network-mediated 710 scholarship transforming traditional scholarly practices * Key 711 information technology trends affecting the conduct of scholarly 712 communication over the next decade * The trend toward end-user computing 713 * The trend toward greater connectivity * Effects of these trends * Key 714 transformations taking place * Summary of principal arguments * 715 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 716 717 Avra MICHELSON, Archival Research and Evaluation Staff, National Archives 718 and Records Administration (NARA), argued that establishing who will use 719 electronic texts and what they will use them for involves a consideration 720 of both information technology and scholarship trends. This 721 consideration includes several elements related to scholarship and 722 technology: 1) the key trends in information technology that are most 723 relevant to scholarship; 2) the key trends in the use of currently 724 available technology by scholars in the nonscientific community; and 3) 725 the relationship between these two very distinct but interrelated trends. 726 The investment in understanding this relationship being made by 727 information providers, technologists, and public policy developers, as 728 well as by scholars themselves, seems to be pervasive and growing, 729 MICHELSON contended. She drew on collaborative work with Jeff Rothenberg 730 on the scholarly use of technology. 731 732 MICHELSON sought to place the phenomenon of electronic texts within the 733 context of broader trends within information technology and scholarly 734 communication. She argued that electronic texts are of most use to 735 researchers to the extent that the researchers' working context (i.e., 736 their relevant bibliographic sources, collegial feedback, analytic tools, 737 notes, drafts, etc.), along with their field's primary and secondary 738 sources, also is accessible in electronic form and can be integrated in 739 ways that are unique to the on-line environment. 740 741 Evaluation of the prospects for the use of electronic texts includes two 742 elements: 1) an examination of the ways in which researchers currently 743 are using electronic texts along with other electronic resources, and 2) 744 an analysis of key information technology trends that are affecting the 745 long-term conduct of scholarly communication. MICHELSON limited her 746 discussion of the use of electronic texts to the practices of humanists 747 and noted that the scientific community was outside the panel's overview. 748 749 MICHELSON examined the nature of the current relationship of electronic 750 texts in particular, and electronic resources in general, to what she 751 maintained were, essentially, five processes of scholarly communication 752 in humanities research. Researchers 1) identify sources, 2) communicate 753 with their colleagues, 3) interpret and analyze data, 4) disseminate 754 their research findings, and 5) prepare curricula to instruct the next 755 generation of scholars and students. This examination would produce a 756 clearer understanding of the synergy among these five processes that 757 fuels the tendency of the use of electronic resources for one process to 758 stimulate its use for other processes of scholarly communication. 759 760 For the first process of scholarly communication, the identification of 761 sources, MICHELSON remarked the opportunity scholars now enjoy to 762 supplement traditional word-of-mouth searches for sources among their 763 colleagues with new forms of electronic searching. So, for example, 764 instead of having to visit the library, researchers are able to explore 765 descriptions of holdings in their offices. Furthermore, if their own 766 institutions' holdings prove insufficient, scholars can access more than 767 200 major American library catalogues over Internet, including the 768 universities of California, Michigan, Pennsylvania, and Wisconsin. 769 Direct access to the bibliographic databases offers intellectual 770 empowerment to scholars by presenting a comprehensive means of browsing 771 through libraries from their homes and offices at their convenience. 772 773 The second process of communication involves communication among 774 scholars. Beyond the most common methods of communication, scholars are 775 using E-mail and a variety of new electronic communications formats 776 derived from it for further academic interchange. E-mail exchanges are 777 growing at an astonishing rate, reportedly 15 percent a month. They 778 currently constitute approximately half the traffic on research and 779 education networks. Moreover, the global spread of E-mail has been so 780 rapid that it is now possible for American scholars to use it to 781 communicate with colleagues in close to 140 other countries. 782 783 Other new exchange formats created by scholars and operating on Internet 784 include more than 700 conferences, with about 80 percent of these devoted 785 to topics in the social sciences and humanities. The rate of growth of 786 these scholarly electronic conferences also is astonishing. From l990 to 787 l991, 200 new conferences were identified on Internet. From October 1991 788 to June 1992, an additional 150 conferences in the social sciences and 789 humanities were added to this directory of listings. Scholars have 790 established conferences in virtually every field, within every different 791 discipline. For example, there are currently close to 600 active social 792 science and humanities conferences on topics such as art and 793 architecture, ethnomusicology, folklore, Japanese culture, medical 794 education, and gifted and talented education. The appeal to scholars of 795 communicating through these conferences is that, unlike any other medium, 796 electronic conferences today provide a forum for global communication 797 with peers at the front end of the research process. 798 799 Interpretation and analysis of sources constitutes the third process of 800 scholarly communication that MICHELSON discussed in terms of texts and 801 textual resources. The methods used to analyze sources fall somewhere on 802 a continuum from quantitative analysis to qualitative analysis. 803 Typically, evidence is culled and evaluated using methods drawn from both 804 ends of this continuum. At one end, quantitative analysis involves the 805 use of mathematical processes such as a count of frequencies and 806 distributions of occurrences or, on a higher level, regression analysis. 807 At the other end of the continuum, qualitative analysis typically 808 involves nonmathematical processes oriented toward language 809 interpretation or the building of theory. Aspects of this work involve 810 the processing--either manual or computational--of large and sometimes 811 massive amounts of textual sources, although the use of nontextual 812 sources as evidence, such as photographs, sound recordings, film footage, 813 and artifacts, is significant as well. 814 815 Scholars have discovered that many of the methods of interpretation and 816 analysis that are related to both quantitative and qualitative methods 817 are processes that can be performed by computers. For example, computers 818 can count. They can count brush strokes used in a Rembrandt painting or 819 perform regression analysis for understanding cause and effect. By means 820 of advanced technologies, computers can recognize patterns, analyze text, 821 and model concepts. Furthermore, computers can complete these processes 822 faster with more sources and with greater precision than scholars who 823 must rely on manual interpretation of data. But if scholars are to use 824 computers for these processes, source materials must be in a form 825 amenable to computer-assisted analysis. For this reason many scholars, 826 once they have identified the sources that are key to their research, are 827 converting them to machine-readable form. Thus, a representative example 828 of the numerous textual conversion projects organized by scholars around 829 the world in recent years to support computational text analysis is the 830 TLG, the Thesaurus Linguae Graecae. This project is devoted to 831 converting the extant ancient texts of classical Greece. (Editor's note: 832 according to the TLG Newsletter of May l992, TLG was in use in thirty-two 833 different countries. This figure updates MICHELSON's previous count by one.) 834 835 The scholars performing these conversions have been asked to recognize 836 that the electronic sources they are converting for one use possess value 837 for other research purposes as well. As a result, during the past few 838 years, humanities scholars have initiated a number of projects to 839 increase scholarly access to converted text. So, for example, the Text 840 Encoding Initiative (TEI), about which more is said later in the program, 841 was established as an effort by scholars to determine standard elements 842 and methods for encoding machine-readable text for electronic exchange. 843 In a second effort to facilitate the sharing of converted text, scholars 844 have created a new institution, the Center for Electronic Texts in the 845 Humanities (CETH). The center estimates that there are 8,000 series of 846 source texts in the humanities that have been converted to 847 machine-readable form worldwide. CETH is undertaking an international 848 search for converted text in the humanities, compiling it into an 849 electronic library, and preparing bibliographic descriptions of the 850 sources for the Research Libraries Information Network's (RLIN) 851 machine-readable data file. The library profession has begun to initiate 852 large conversion projects as well, such as American Memory. 853 854 While scholars have been making converted text available to one another, 855 typically on disk or on CD-ROM, the clear trend is toward making these 856 resources available through research and education networks. Thus, the 857 American and French Research on the Treasury of the French Language 858 (ARTFL) and the Dante Project are already available on Internet. 859 MICHELSON summarized this section on interpretation and analysis by 860 noting that: 1) increasing numbers of humanities scholars in the library 861 community are recognizing the importance to the advancement of 862 scholarship of retrospective conversion of source materials in the arts 863 and humanities; and 2) there is a growing realization that making the 864 sources available on research and education networks maximizes their 865 usefulness for the analysis performed by humanities scholars. 866 867 The fourth process of scholarly communication is dissemination of 868 research findings, that is, publication. Scholars are using existing 869 research and education networks to engineer a new type of publication: 870 scholarly-controlled journals that are electronically produced and 871 disseminated. Although such journals are still emerging as a 872 communication format, their number has grown, from approximately twelve 873 to thirty-six during the past year (July 1991 to June 1992). Most of 874 these electronic scholarly journals are devoted to topics in the 875 humanities. As with network conferences, scholarly enthusiasm for these 876 electronic journals stems from the medium's unique ability to advance 877 scholarship in a way that no other medium can do by supporting global 878 feedback and interchange, practically in real time, early in the research 879 process. Beyond scholarly journals, MICHELSON remarked the delivery of 880 commercial full-text products, such as articles in professional journals, 881 newsletters, magazines, wire services, and reference sources. These are 882 being delivered via on-line local library catalogues, especially through 883 CD-ROMs. Furthermore, according to MICHELSON, there is general optimism 884 that the copyright and fees issues impeding the delivery of full text on 885 existing research and education networks soon will be resolved. 886 887 The final process of scholarly communication is curriculum development 888 and instruction, and this involves the use of computer information 889 technologies in two areas. The first is the development of 890 computer-oriented instructional tools, which includes simulations, 891 multimedia applications, and computer tools that are used to assist in 892 the analysis of sources in the classroom, etc. The Perseus Project, a 893 database that provides a multimedia curriculum on classical Greek 894 civilization, is a good example of the way in which entire curricula are 895 being recast using information technologies. It is anticipated that the 896 current difficulty in exchanging electronically computer-based 897 instructional software, which in turn makes it difficult for one scholar 898 to build upon the work of others, will be resolved before too long. 899 Stand-alone curricular applications that involve electronic text will be 900 sharable through networks, reinforcing their significance as intellectual 901 products as well as instructional tools. 902 903 The second aspect of electronic learning involves the use of research and 904 education networks for distance education programs. Such programs 905 interactively link teachers with students in geographically scattered 906 locations and rely on the availability of electronic instructional 907 resources. Distance education programs are gaining wide appeal among 908 state departments of education because of their demonstrated capacity to 909 bring advanced specialized course work and an array of experts to many 910 classrooms. A recent report found that at least 32 states operated at 911 least one statewide network for education in 1991, with networks under 912 development in many of the remaining states. 913 914 MICHELSON summarized this section by noting two striking changes taking 915 place in scholarly communication among humanities scholars. First is the 916 extent to which electronic text in particular, and electronic resources 917 in general, are being infused into each of the five processes described 918 above. As mentioned earlier, there is a certain synergy at work here. 919 The use of electronic resources for one process tends to stimulate its 920 use for other processes, because the chief course of movement is toward a 921 comprehensive on-line working context for humanities scholars that 922 includes on-line availability of key bibliographies, scholarly feedback, 923 sources, analytical tools, and publications. MICHELSON noted further 924 that the movement toward a comprehensive on-line working context for 925 humanities scholars is not new. In fact, it has been underway for more 926 than forty years in the humanities, since Father Roberto Busa began 927 developing an electronic concordance of the works of Saint Thomas Aquinas 928 in 1949. What we are witnessing today, MICHELSON contended, is not the 929 beginning of this on-line transition but, for at least some humanities 930 scholars, the turning point in the transition from a print to an 931 electronic working context. Coinciding with the on-line transition, the 932 second striking change is the extent to which research and education 933 networks are becoming the new medium of scholarly communication. The 934 existing Internet and the pending National Education and Research Network 935 (NREN) represent the new meeting ground where scholars are going for 936 bibliographic information, scholarly dialogue and feedback, the most 937 current publications in their field, and high-level educational 938 offerings. Traditional scholarly practices are undergoing tremendous 939 transformations as a result of the emergence and growing prominence of 940 what is called network-mediated scholarship. 941 942 MICHELSON next turned to the second element of the framework she proposed 943 at the outset of her talk for evaluating the prospects for electronic 944 text, namely the key information technology trends affecting the conduct 945 of scholarly communication over the next decade: 1) end-user computing 946 and 2) connectivity. 947 948 End-user computing means that the person touching the keyboard, or 949 performing computations, is the same as the person who initiates or 950 consumes the computation. The emergence of personal computers, along 951 with a host of other forces, such as ubiquitous computing, advances in 952 interface design, and the on-line transition, is prompting the consumers 953 of computation to do their own computing, and is thus rendering obsolete 954 the traditional distinction between end users and ultimate users. 955 956 The trend toward end-user computing is significant to consideration of 957 the prospects for electronic texts because it means that researchers are 958 becoming more adept at doing their own computations and, thus, more 959 competent in the use of electronic media. By avoiding programmer 960 intermediaries, computation is becoming central to the researcher's 961 thought process. This direct involvement in computing is changing the 962 researcher's perspective on the nature of research itself, that is, the 963 kinds of questions that can be posed, the analytical methodologies that 964 can be used, the types and amount of sources that are appropriate for 965 analyses, and the form in which findings are presented. The trend toward 966 end-user computing means that, increasingly, electronic media and 967 computation are being infused into all processes of humanities 968 scholarship, inspiring remarkable transformations in scholarly 969 communication. 970 971 The trend toward greater connectivity suggests that researchers are using 972 computation increasingly in network environments. Connectivity is 973 important to scholarship because it erases the distance that separates 974 students from teachers and scholars from their colleagues, while allowing 975 users to access remote databases, share information in many different 976 media, connect to their working context wherever they are, and 977 collaborate in all phases of research. 978 979 The combination of the trend toward end-user computing and the trend 980 toward connectivity suggests that the scholarly use of electronic 981 resources, already evident among some researchers, will soon become an 982 established feature of scholarship. The effects of these trends, along 983 with ongoing changes in scholarly practices, point to a future in which 984 humanities researchers will use computation and electronic communication 985 to help them formulate ideas, access sources, perform research, 986 collaborate with colleagues, seek peer review, publish and disseminate 987 results, and engage in many other professional and educational activities. 988 989 In summary, MICHELSON emphasized four points: 1) A portion of humanities 990 scholars already consider electronic texts the preferred format for 991 analysis and dissemination. 2) Scholars are using these electronic 992 texts, in conjunction with other electronic resources, in all the 993 processes of scholarly communication. 3) The humanities scholars' 994 working context is in the process of changing from print technology to 995 electronic technology, in many ways mirroring transformations that have 996 occurred or are occurring within the scientific community. 4) These 997 changes are occurring in conjunction with the development of a new 998 communication medium: research and education networks that are 999 characterized by their capacity to advance scholarship in a wholly unique 1000 way. 1001 1002 MICHELSON also reiterated her three principal arguments: l) Electronic 1003 texts are best understood in terms of the relationship to other 1004 electronic resources and the growing prominence of network-mediated 1005 scholarship. 2) The prospects for electronic texts lie in their capacity 1006 to be integrated into the on-line network of electronic resources that 1007 comprise the new working context for scholars. 3) Retrospective conversion 1008 of portions of the scholarly record should be a key strategy as information 1009 providers respond to changes in scholarly communication practices. 1010 1011 ****** 1012 1013 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1014 VECCIA * AM's evaluation project and public users of electronic resources 1015 * AM and its design * Site selection and evaluating the Macintosh 1016 implementation of AM * Characteristics of the six public libraries 1017 selected * Characteristics of AM's users in these libraries * Principal 1018 ways AM is being used * 1019 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1020 1021 Susan VECCIA, team leader, and Joanne FREEMAN, associate coordinator, 1022 American Memory, Library of Congress, gave a joint presentation. First, 1023 by way of introduction, VECCIA explained her and FREEMAN's roles in 1024 American Memory (AM). Serving principally as an observer, VECCIA has 1025 assisted with the evaluation project of AM, placing AM collections in a 1026 variety of different sites around the country and helping to organize and 1027 implement that project. FREEMAN has been an associate coordinator of AM 1028 and has been involved principally with the interpretative materials, 1029 preparing some of the electronic exhibits and printed historical 1030 information that accompanies AM and that is requested by users. VECCIA 1031 and FREEMAN shared anecdotal observations concerning AM with public users 1032 of electronic resources. Notwithstanding a fairly structured evaluation 1033 in progress, both VECCIA and FREEMAN chose not to report on specifics in 1034 terms of numbers, etc., because they felt it was too early in the 1035 evaluation project to do so. 1036 1037 AM is an electronic archive of primary source materials from the Library 1038 of Congress, selected collections representing a variety of formats-- 1039 photographs, graphic arts, recorded sound, motion pictures, broadsides, 1040 and soon, pamphlets and books. In terms of the design of this system, 1041 the interpretative exhibits have been kept separate from the primary 1042 resources, with good reason. Accompanying this collection are printed 1043 documentation and user guides, as well as guides that FREEMAN prepared for 1044 teachers so that they may begin using the content of the system at once. 1045 1046 VECCIA described the evaluation project before talking about the public 1047 users of AM, limiting her remarks to public libraries, because FREEMAN 1048 would talk more specifically about schools from kindergarten to twelfth 1049 grade (K-12). Having started in spring 1991, the evaluation currently 1050 involves testing of the Macintosh implementation of AM. Since the 1051 primary goal of this evaluation is to determine the most appropriate 1052 audience or audiences for AM, very different sites were selected. This 1053 makes evaluation difficult because of the varying degrees of technology 1054 literacy among the sites. AM is situated in forty-four locations, of 1055 which six are public libraries and sixteen are schools. Represented 1056 among the schools are elementary, junior high, and high schools. 1057 District offices also are involved in the evaluation, which will 1058 conclude in summer 1993. 1059 1060 VECCIA focused the remainder of her talk on the six public libraries, one 1061 of which doubles as a state library. They represent a range of 1062 geographic areas and a range of demographic characteristics. For 1063 example, three are located in urban settings, two in rural settings, and 1064 one in a suburban setting. A range of technical expertise is to be found 1065 among these facilities as well. For example, one is an "Apple library of 1066 the future," while two others are rural one-room libraries--in one, AM 1067 sits at the front desk next to a tractor manual. 1068 1069 All public libraries have been extremely enthusiastic, supportive, and 1070 appreciative of the work that AM has been doing. VECCIA characterized 1071 various users: Most users in public libraries describe themselves as 1072 general readers; of the students who use AM in the public libraries, 1073 those in fourth grade and above seem most interested. Public libraries 1074 in rural sites tend to attract retired people, who have been highly 1075 receptive to AM. Users tend to fall into two additional categories: 1076 people interested in the content and historical connotations of these 1077 primary resources, and those fascinated by the technology. The format 1078 receiving the most comments has been motion pictures. The adult users in 1079 public libraries are more comfortable with IBM computers, whereas young 1080 people seem comfortable with either IBM or Macintosh, although most of 1081 them seem to come from a Macintosh background. This same tendency is 1082 found in the schools. 1083 1084 What kinds of things do users do with AM? In a public library there are 1085 two main goals or ways that AM is being used: as an individual learning 1086 tool, and as a leisure activity. Adult learning was one area that VECCIA 1087 would highlight as a possible application for a tool such as AM. She 1088 described a patron of a rural public library who comes in every day on 1089 his lunch hour and literally reads AM, methodically going through the 1090 collection image by image. At the end of his hour he makes an electronic 1091 bookmark, puts it in his pocket, and returns to work. The next day he 1092 comes in and resumes where he left off. Interestingly, this man had 1093 never been in the library before he used AM. In another small, rural 1094 library, the coordinator reports that AM is a popular activity for some 1095 of the older, retired people in the community, who ordinarily would not 1096 use "those things,"--computers. Another example of adult learning in 1097 public libraries is book groups, one of which, in particular, is using AM 1098 as part of its reading on industrialization, integration, and urbanization 1099 in the early 1900s. 1100 1101 One library reports that a family is using AM to help educate their 1102 children. In another instance, individuals from a local museum came in 1103 to use AM to prepare an exhibit on toys of the past. These two examples 1104 emphasize the mission of the public library as a cultural institution, 1105 reaching out to people who do not have the same resources available to 1106 those who live in a metropolitan area or have access to a major library. 1107 One rural library reports that junior high school students in large 1108 numbers came in one afternoon to use AM for entertainment. A number of 1109 public libraries reported great interest among postcard collectors in the 1110 Detroit collection, which was essentially a collection of images used on 1111 postcards around the turn of the century. Train buffs are similarly 1112 interested because that was a time of great interest in railroading. 1113 People, it was found, relate to things that they know of firsthand. For 1114 example, in both rural public libraries where AM was made available, 1115 observers reported that the older people with personal remembrances of 1116 the turn of the century were gravitating to the Detroit collection. 1117 These examples served to underscore MICHELSON's observation re the 1118 integration of electronic tools and ideas--that people learn best when 1119 the material relates to something they know. 1120 1121 VECCIA made the final point that in many cases AM serves as a 1122 public-relations tool for the public libraries that are testing it. In 1123 one case, AM is being used as a vehicle to secure additional funding for 1124 the library. In another case, AM has served as an inspiration to the 1125 staff of a major local public library in the South to think about ways to 1126 make its own collection of photographs more accessible to the public. 1127 1128 ****** 1129 1130 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1131 FREEMAN * AM and archival electronic resources in a school environment * 1132 Questions concerning context * Questions concerning the electronic format 1133 itself * Computer anxiety * Access and availability of the system * 1134 Hardware * Strengths gained through the use of archival resources in 1135 schools * 1136 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1137 1138 Reiterating an observation made by VECCIA, that AM is an archival 1139 resource made up of primary materials with very little interpretation, 1140 FREEMAN stated that the project has attempted to bridge the gap between 1141 these bare primary materials and a school environment, and in that cause 1142 has created guided introductions to AM collections. Loud demand from the 1143 educational community, chiefly from teachers working with the upper 1144 grades of elementary school through high school, greeted the announcement 1145 that AM would be tested around the country. 1146 1147 FREEMAN reported not only on what was learned about AM in a school 1148 environment, but also on several universal questions that were raised 1149 concerning archival electronic resources in schools. She discussed 1150 several strengths of this type of material in a school environment as 1151 opposed to a highly structured resource that offers a limited number of 1152 paths to follow. 1153 1154 FREEMAN first raised several questions about using AM in a school 1155 environment. There is often some difficulty in developing a sense of 1156 what the system contains. Many students sit down at a computer resource 1157 and assume that, because AM comes from the Library of Congress, all of 1158 American history is now at their fingertips. As a result of that sort of 1159 mistaken judgment, some students are known to conclude that AM contains 1160 nothing of use to them when they look for one or two things and do not 1161 find them. It is difficult to discover that middle ground where one has 1162 a sense of what the system contains. Some students grope toward the idea 1163 of an archive, a new idea to them, since they have not previously 1164 experienced what it means to have access to a vast body of somewhat 1165 random information. 1166 1167 Other questions raised by FREEMAN concerned the electronic format itself. 1168 For instance, in a school environment it is often difficult both for 1169 teachers and students to gain a sense of what it is they are viewing. 1170 They understand that it is a visual image, but they do not necessarily 1171 know that it is a postcard from the turn of the century, a panoramic 1172 photograph, or even machine-readable text of an eighteenth-century 1173 broadside, a twentieth-century printed book, or a nineteenth-century 1174 diary. That distinction is often difficult for people in a school 1175 environment to grasp. Because of that, it occasionally becomes difficult 1176 to draw conclusions from what one is viewing. 1177 1178 FREEMAN also noted the obvious fear of the computer, which constitutes a 1179 difficulty in using an electronic resource. Though students in general 1180 did not suffer from this anxiety, several older students feared that they 1181 were computer-illiterate, an assumption that became self-fulfilling when 1182 they searched for something but failed to find it. FREEMAN said she 1183 believed that some teachers also fear computer resources, because they 1184 believe they lack complete control. FREEMAN related the example of 1185 teachers shooing away students because it was not their time to use the 1186 system. This was a case in which the situation had to be extremely 1187 structured so that the teachers would not feel that they had lost their 1188 grasp on what the system contained. 1189 1190 A final question raised by FREEMAN concerned access and availability of 1191 the system. She noted the occasional existence of a gap in communication 1192 between school librarians and teachers. Often AM sits in a school 1193 library and the librarian is the person responsible for monitoring the 1194 system. Teachers do not always take into their world new library 1195 resources about which the librarian is excited. Indeed, at the sites 1196 where AM had been used most effectively within a library, the librarian 1197 was required to go to specific teachers and instruct them in its use. As 1198 a result, several AM sites will have in-service sessions over a summer, 1199 in the hope that perhaps, with a more individualized link, teachers will 1200 be more likely to use the resource. 1201 1202 A related issue in the school context concerned the number of 1203 workstations available at any one location. Centralization of equipment 1204 at the district level, with teachers invited to download things and walk 1205 away with them, proved unsuccessful because the hours these offices were 1206 open were also school hours. 1207 1208 Another issue was hardware. As VECCIA observed, a range of sites exists, 1209 some technologically advanced and others essentially acquiring their 1210 first computer for the primary purpose of using it in conjunction with 1211 AM's testing. Users at technologically sophisticated sites want even 1212 more sophisticated hardware, so that they can perform even more 1213 sophisticated tasks with the materials in AM. But once they acquire a 1214 newer piece of hardware, they must learn how to use that also; at an 1215 unsophisticated site it takes an extremely long time simply to become 1216 accustomed to the computer, not to mention the program offered with the 1217 computer. All of these small issues raise one large question, namely, 1218 are systems like AM truly rewarding in a school environment, or do they 1219 simply act as innovative toys that do little more than spark interest? 1220 1221 FREEMAN contended that the evaluation project has revealed several strengths 1222 that were gained through the use of archival resources in schools, including: 1223 1224 * Psychic rewards from using AM as a vast, rich database, with 1225 teachers assigning various projects to students--oral presentations, 1226 written reports, a documentary, a turn-of-the-century newspaper-- 1227 projects that start with the materials in AM but are completed using 1228 other resources; AM thus is used as a research tool in conjunction 1229 with other electronic resources, as well as with books and items in 1230 the library where the system is set up. 1231 1232 * Students are acquiring computer literacy in a humanities context. 1233 1234 * This sort of system is overcoming the isolation between disciplines 1235 that often exists in schools. For example, many English teachers are 1236 requiring their students to write papers on historical topics 1237 represented in AM. Numerous teachers have reported that their 1238 students are learning critical thinking skills using the system. 1239 1240 * On a broader level, AM is introducing primary materials, not only 1241 to students but also to teachers, in an environment where often 1242 simply none exist--an exciting thing for the students because it 1243 helps them learn to conduct research, to interpret, and to draw 1244 their own conclusions. In learning to conduct research and what it 1245 means, students are motivated to seek knowledge. That relates to 1246 another positive outcome--a high level of personal involvement of 1247 students with the materials in this system and greater motivation to 1248 conduct their own research and draw their own conclusions. 1249 1250 * Perhaps the most ironic strength of these kinds of archival 1251 electronic resources is that many of the teachers AM interviewed 1252 were desperate, it is no exaggeration to say, not only for primary 1253 materials but for unstructured primary materials. These would, they 1254 thought, foster personally motivated research, exploration, and 1255 excitement in their students. Indeed, these materials have done 1256 just that. Ironically, however, this lack of structure produces 1257 some of the confusion to which the newness of these kinds of 1258 resources may also contribute. The key to effective use of archival 1259 products in a school environment is a clear, effective introduction 1260 to the system and to what it contains. 1261 1262 ****** 1263 1264 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1265 DISCUSSION * Nothing known, quantitatively, about the number of 1266 humanities scholars who must see the original versus those who would 1267 settle for an edited transcript, or about the ways in which humanities 1268 scholars are using information technology * Firm conclusions concerning 1269 the manner and extent of the use of supporting materials in print 1270 provided by AM to await completion of evaluative study * A listener's 1271 reflections on additional applications of electronic texts * Role of 1272 electronic resources in teaching elementary research skills to students * 1273 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1274 1275 During the discussion that followed the presentations by MICHELSON, 1276 VECCIA, and FREEMAN, additional points emerged. 1277 1278 LESK asked if MICHELSON could give any quantitative estimate of the 1279 number of humanities scholars who must see or want to see the original, 1280 or the best possible version of the material, versus those who typically 1281 would settle for an edited transcript. While unable to provide a figure, 1282 she offered her impressions as an archivist who has done some reference 1283 work and has discussed this issue with other archivists who perform 1284 reference, that those who use archives and those who use primary sources 1285 for what would be considered very high-level scholarly research, as 1286 opposed to, say, undergraduate papers, were few in number, especially 1287 given the public interest in using primary sources to conduct 1288 genealogical or avocational research and the kind of professional 1289 research done by people in private industry or the federal government. 1290 More important in MICHELSON's view was that, quantitatively, nothing is 1291 known about the ways in which, for example, humanities scholars are using 1292 information technology. No studies exist to offer guidance in creating 1293 strategies. The most recent study was conducted in 1985 by the American 1294 Council of Learned Societies (ACLS), and what it showed was that 50 1295 percent of humanities scholars at that time were using computers. That 1296 constitutes the extent of our knowledge. 1297 1298 Concerning AM's strategy for orienting people toward the scope of 1299 electronic resources, FREEMAN could offer no hard conclusions at this 1300 point, because she and her colleagues were still waiting to see, 1301 particularly in the schools, what has been made of their efforts. Within 1302 the system, however, AM has provided what are called electronic exhibits- 1303 -such as introductions to time periods and materials--and these are 1304 intended to offer a student user a sense of what a broadside is and what 1305 it might tell her or him. But FREEMAN conceded that the project staff 1306 would have to talk with students next year, after teachers have had a 1307 summer to use the materials, and attempt to discover what the students 1308 were learning from the materials. In addition, FREEMAN described 1309 supporting materials in print provided by AM at the request of local 1310 teachers during a meeting held at LC. These included time lines, 1311 bibliographies, and other materials that could be reproduced on a 1312 photocopier in a classroom. Teachers could walk away with and use these, 1313 and in this way gain a better understanding of the contents. But again, 1314 reaching firm conclusions concerning the manner and extent of their use 1315 would have to wait until next year. 1316 1317 As to the changes she saw occurring at the National Archives and Records 1318 Administration (NARA) as a result of the increasing emphasis on 1319 technology in scholarly research, MICHELSON stated that NARA at this 1320 point was absorbing the report by her and Jeff Rothenberg addressing 1321 strategies for the archival profession in general, although not for the 1322 National Archives specifically. NARA is just beginning to establish its 1323 role and what it can do. In terms of changes and initiatives that NARA 1324 can take, no clear response could be given at this time. 1325 1326 GREENFIELD remarked two trends mentioned in the session. Reflecting on 1327 DALY's opening comments on how he could have used a Latin collection of 1328 text in an electronic form, he said that at first he thought most scholars 1329 would be unwilling to do that. But as he thought of that in terms of the 1330 original meaning of research--that is, having already mastered these texts, 1331 researching them for critical and comparative purposes--for the first time, 1332 the electronic format made a lot of sense. GREENFIELD could envision 1333 growing numbers of scholars learning the new technologies for that very 1334 aspect of their scholarship and for convenience's sake. 1335 1336 Listening to VECCIA and FREEMAN, GREENFIELD thought of an additional 1337 application of electronic texts. He realized that AM could be used as a 1338 guide to lead someone to original sources. Students cannot be expected 1339 to have mastered these sources, things they have never known about 1340 before. Thus, AM is leading them, in theory, to a vast body of 1341 information and giving them a superficial overview of it, enabling them 1342 to select parts of it. GREENFIELD asked if any evidence exists that this 1343 resource will indeed teach the new user, the K-12 students, how to do 1344 research. Scholars already know how to do research and are applying 1345 these new tools. But he wondered why students would go beyond picking 1346 out things that were most exciting to them. 1347 1348 FREEMAN conceded the correctness of GREENFIELD's observation as applied 1349 to a school environment. The risk is that a student would sit down at a 1350 system, play with it, find some things of interest, and then walk away. 1351 But in the relatively controlled situation of a school library, much will 1352 depend on the instructions a teacher or a librarian gives a student. She 1353 viewed the situation not as one of fine-tuning research skills but of 1354 involving students at a personal level in understanding and researching 1355 things. Given the guidance one can receive at school, it then becomes 1356 possible to teach elementary research skills to students, which in fact 1357 one particular librarian said she was teaching her fifth graders. 1358 FREEMAN concluded that introducing the idea of following one's own path 1359 of inquiry, which is essentially what research entails, involves more 1360 than teaching specific skills. To these comments VECCIA added the 1361 observation that the individual teacher and the use of a creative 1362 resource, rather than AM itself, seemed to make the key difference. 1363 Some schools and some teachers are making excellent use of the nature 1364 of critical thinking and teaching skills, she said. 1365 1366 Concurring with these remarks, DALY closed the session with the thought that 1367 the more that producers produced for teachers and for scholars to use with 1368 their students, the more successful their electronic products would prove. 1369 1370 ****** 1371 1372 SESSION II. SHOW AND TELL 1373 1374 Jacqueline HESS, director, National Demonstration Laboratory, served as 1375 moderator of the "show-and-tell" session. She noted that a 1376 question-and-answer period would follow each presentation. 1377 1378 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1379 MYLONAS * Overview and content of Perseus * Perseus' primary materials 1380 exist in a system-independent, archival form * A concession * Textual 1381 aspects of Perseus * Tools to use with the Greek text * Prepared indices 1382 and full-text searches in Perseus * English-Greek word search leads to 1383 close study of words and concepts * Navigating Perseus by tracing down 1384 indices * Using the iconography to perform research * 1385 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1386 1387 Elli MYLONAS, managing editor, Perseus Project, Harvard University, first 1388 gave an overview of Perseus, a large, collaborative effort based at 1389 Harvard University but with contributors and collaborators located at 1390 numerous universities and colleges in the United States (e.g., Bowdoin, 1391 Maryland, Pomona, Chicago, Virginia). Funded primarily by the 1392 Annenberg/CPB Project, with additional funding from Apple, Harvard, and 1393 the Packard Humanities Institute, among others, Perseus is a multimedia, 1394 hypertextual database for teaching and research on classical Greek 1395 civilization, which was released in February 1992 in version 1.0 and 1396 distributed by Yale University Press. 1397 1398 Consisting entirely of primary materials, Perseus includes ancient Greek 1399 texts and translations of those texts; catalog entries--that is, museum 1400 catalog entries, not library catalog entries--on vases, sites, coins, 1401 sculpture, and archaeological objects; maps; and a dictionary, among 1402 other sources. The number of objects and the objects for which catalog 1403 entries exist are accompanied by thousands of color images, which 1404 constitute a major feature of the database. Perseus contains 1405 approximately 30 megabytes of text, an amount that will double in 1406 subsequent versions. In addition to these primary materials, the Perseus 1407 Project has been building tools for using them, making access and 1408 navigation easier, the goal being to build part of the electronic 1409 environment discussed earlier in the morning in which students or 1410 scholars can work with their sources. 1411 1412 The demonstration of Perseus will show only a fraction of the real work 1413 that has gone into it, because the project had to face the dilemma of 1414 what to enter when putting something into machine-readable form: should 1415 one aim for very high quality or make concessions in order to get the 1416 material in? Since Perseus decided to opt for very high quality, all of 1417 its primary materials exist in a system-independent--insofar as it is 1418 possible to be system-independent--archival form. Deciding what that 1419 archival form would be and attaining it required much work and thought. 1420 For example, all the texts are marked up in SGML, which will be made 1421 compatible with the guidelines of the Text Encoding Initiative (TEI) when 1422 they are issued. 1423 1424 Drawings are postscript files, not meeting international standards, but 1425 at least designed to go across platforms. Images, or rather the real 1426 archival forms, consist of the best available slides, which are being 1427 digitized. Much of the catalog material exists in database form--a form 1428 that the average user could use, manipulate, and display on a personal 1429 computer, but only at great cost. Thus, this is where the concession 1430 comes in: All of this rich, well-marked-up information is stripped of 1431 much of its content; the images are converted into bit-maps and the text 1432 into small formatted chunks. All this information can then be imported 1433 into HyperCard and run on a mid-range Macintosh, which is what Perseus 1434 users have. This fact has made it possible for Perseus to attain wide 1435 use fairly rapidly. Without those archival forms the HyperCard version 1436 being demonstrated could not be made easily, and the project could not 1437 have the potential to move to other forms and machines and software as 1438 they appear, none of which information is in Perseus on the CD. 1439 1440 Of the numerous multimedia aspects of Perseus, MYLONAS focused on the 1441 textual. Part of what makes Perseus such a pleasure to use, MYLONAS 1442 said, is this effort at seamless integration and the ability to move 1443 around both visual and textual material. Perseus also made the decision 1444 not to attempt to interpret its material any more than one interprets by 1445 selecting. But, MYLONAS emphasized, Perseus is not courseware: No 1446 syllabus exists. There is no effort to define how one teaches a topic 1447 using Perseus, although the project may eventually collect papers by 1448 people who have used it to teach. Rather, Perseus aims to provide 1449 primary material in a kind of electronic library, an electronic sandbox, 1450 so to say, in which students and scholars who are working on this 1451 material can explore by themselves. With that, MYLONAS demonstrated 1452 Perseus, beginning with the Perseus gateway, the first thing one sees 1453 upon opening Perseus--an effort in part to solve the contextualizing 1454 problem--which tells the user what the system contains. 1455 1456 MYLONAS demonstrated only a very small portion, beginning with primary 1457 texts and running off the CD-ROM. Having selected Aeschylus' Prometheus 1458 Bound, which was viewable in Greek and English pretty much in the same 1459 segments together, MYLONAS demonstrated tools to use with the Greek text, 1460 something not possible with a book: looking up the dictionary entry form 1461 of an unfamiliar word in Greek after subjecting it to Perseus' 1462 morphological analysis for all the texts. After finding out about a 1463 word, a user may then decide to see if it is used anywhere else in Greek. 1464 Because vast amounts of indexing support all of the primary material, one 1465 can find out where else all forms of a particular Greek word appear-- 1466 often not a trivial matter because Greek is highly inflected. Further, 1467 since the story of Prometheus has to do with the origins of sacrifice, a 1468 user may wish to study and explore sacrifice in Greek literature; by 1469 typing sacrifice into a small window, a user goes to the English-Greek 1470 word list--something one cannot do without the computer (Perseus has 1471 indexed the definitions of its dictionary)--the string sacrifice appears 1472 in the definitions of these sixty-five words. One may then find out 1473 where any of those words is used in the work(s) of a particular author. 1474 The English definitions are not lemmatized. 1475 1476 All of the indices driving this kind of usage were originally devised for 1477 speed, MYLONAS observed; in other words, all that kind of information-- 1478 all forms of all words, where they exist, the dictionary form they belong 1479 to--were collected into databases, which will expedite searching. Then 1480 it was discovered that one can do things searching in these databases 1481 that could not be done searching in the full texts. Thus, although there 1482 are full-text searches in Perseus, much of the work is done behind the 1483 scenes, using prepared indices. Re the indexing that is done behind the 1484 scenes, MYLONAS pointed out that without the SGML forms of the text, it 1485 could not be done effectively. Much of this indexing is based on the 1486 structures that are made explicit by the SGML tagging. 1487 1488 It was found that one of the things many of Perseus' non-Greek-reading 1489 users do is start from the dictionary and then move into the close study 1490 of words and concepts via this kind of English-Greek word search, by which 1491 means they might select a concept. This exercise has been assigned to 1492 students in core courses at Harvard--to study a concept by looking for the 1493 English word in the dictionary, finding the Greek words, and then finding 1494 the words in the Greek but, of course, reading across in the English. 1495 That tells them a great deal about what a translation means as well. 1496 1497 Should one also wish to see images that have to do with sacrifice, that 1498 person would go to the object key word search, which allows one to 1499 perform a similar kind of index retrieval on the database of 1500 archaeological objects. Without words, pictures are useless; Perseus has 1501 not reached the point where it can do much with images that are not 1502 cataloged. Thus, although it is possible in Perseus with text and images 1503 to navigate by knowing where one wants to end up--for example, a 1504 red-figure vase from the Boston Museum of Fine Arts--one can perform this 1505 kind of navigation very easily by tracing down indices. MYLONAS 1506 illustrated several generic scenes of sacrifice on vases. The features 1507 demonstrated derived from Perseus 1.0; version 2.0 will implement even 1508 better means of retrieval. 1509 1510 MYLONAS closed by looking at one of the pictures and noting again that 1511 one can do a great deal of research using the iconography as well as the 1512 texts. For instance, students in a core course at Harvard this year were 1513 highly interested in Greek concepts of foreigners and representations of 1514 non-Greeks. So they performed a great deal of research, both with texts 1515 (e.g., Herodotus) and with iconography on vases and coins, on how the 1516 Greeks portrayed non-Greeks. At the same time, art historians who study 1517 iconography were also interested, and were able to use this material. 1518 1519 ****** 1520 1521 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1522 DISCUSSION * Indexing and searchability of all English words in Perseus * 1523 Several features of Perseus 1.0 * Several levels of customization 1524 possible * Perseus used for general education * Perseus' effects on 1525 education * Contextual information in Perseus * Main challenge and 1526 emphasis of Perseus * 1527 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1528 1529 Several points emerged in the discussion that followed MYLONAS's presentation. 1530 1531 Although MYLONAS had not demonstrated Perseus' ability to cross-search 1532 documents, she confirmed that all English words in Perseus are indexed 1533 and can be searched. So, for example, sacrifice could have been searched 1534 in all texts, the historical essay, and all the catalogue entries with 1535 their descriptions--in short, in all of Perseus. 1536 1537 Boolean logic is not in Perseus 1.0 but will be added to the next 1538 version, although an effort is being made not to restrict Perseus to a 1539 database in which one just performs searching, Boolean or otherwise. It 1540 is possible to move laterally through the documents by selecting a word 1541 one is interested in and selecting an area of information one is 1542 interested in and trying to look that word up in that area. 1543 1544 Since Perseus was developed in HyperCard, several levels of customization 1545 are possible. Simple authoring tools exist that allow one to create 1546 annotated paths through the information, which are useful for note-taking 1547 and for guided tours for teaching purposes and for expository writing. 1548 With a little more ingenuity it is possible to begin to add or substitute 1549 material in Perseus. 1550 1551 Perseus has not been used so much for classics education as for general 1552 education, where it seemed to have an impact on the students in the core 1553 course at Harvard (a general required course that students must take in 1554 certain areas). Students were able to use primary material much more. 1555 1556 The Perseus Project has an evaluation team at the University of Maryland 1557 that has been documenting Perseus' effects on education. Perseus is very 1558 popular, and anecdotal evidence indicates that it is having an effect at 1559 places other than Harvard, for example, test sites at Ball State 1560 University, Drury College, and numerous small places where opportunities 1561 to use vast amounts of primary data may not exist. One documented effect 1562 is that archaeological, anthropological, and philological research is 1563 being done by the same person instead of by three different people. 1564 1565 The contextual information in Perseus includes an overview essay, a 1566 fairly linear historical essay on the fifth century B.C. that provides 1567 links into the primary material (e.g., Herodotus, Thucydides, and 1568 Plutarch), via small gray underscoring (on the screen) of linked 1569 passages. These are handmade links into other material. 1570 1571 To different extents, most of the production work was done at Harvard, 1572 where the people and the equipment are located. Much of the 1573 collaborative activity involved data collection and structuring, because 1574 the main challenge and the emphasis of Perseus is the gathering of 1575 primary material, that is, building a useful environment for studying 1576 classical Greece, collecting data, and making it useful. 1577 Systems-building is definitely not the main concern. Thus, much of the 1578 work has involved writing essays, collecting information, rewriting it, 1579 and tagging it. That can be done off site. The creative link for the 1580 overview essay as well as for both systems and data was collaborative, 1581 and was forged via E-mail and paper mail with professors at Pomona and 1582 Bowdoin. 1583 1584 ****** 1585 1586 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1587 CALALUCA * PLD's principal focus and contribution to scholarship * 1588 Various questions preparatory to beginning the project * Basis for 1589 project * Basic rule in converting PLD * Concerning the images in PLD * 1590 Running PLD under a variety of retrieval softwares * Encoding the 1591 database a hard-fought issue * Various features demonstrated * Importance 1592 of user documentation * Limitations of the CD-ROM version * 1593 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1594 1595 Eric CALALUCA, vice president, Chadwyck-Healey, Inc., demonstrated a 1596 software interpretation of the Patrologia Latina Database (PLD). PLD's 1597 principal focus from the beginning of the project about three-and-a-half 1598 years ago was on converting Migne's Latin series, and in the end, 1599 CALALUCA suggested, conversion of the text will be the major contribution 1600 to scholarship. CALALUCA stressed that, as possibly the only private 1601 publishing organization at the Workshop, Chadwyck-Healey had sought no 1602 federal funds or national foundation support before embarking upon the 1603 project, but instead had relied upon a great deal of homework and 1604 marketing to accomplish the task of conversion. 1605 1606 Ever since the possibilities of computer-searching have emerged, scholars 1607 in the field of late ancient and early medieval studies (philosophers, 1608 theologians, classicists, and those studying the history of natural law 1609 and the history of the legal development of Western civilization) have 1610 been longing for a fully searchable version of Western literature, for 1611 example, all the texts of Augustine and Bernard of Clairvaux and 1612 Boethius, not to mention all the secondary and tertiary authors. 1613 1614 Various questions arose, CALALUCA said. Should one convert Migne? 1615 Should the database be encoded? Is it necessary to do that? How should 1616 it be delivered? What about CD-ROM? Since this is a transitional 1617 medium, why even bother to create software to run on a CD-ROM? Since 1618 everybody knows people will be networking information, why go to the 1619 trouble--which is far greater with CD-ROM than with the production of 1620 magnetic data? Finally, how does one make the data available? Can many 1621 of the hurdles to using electronic information that some publishers have 1622 imposed upon databases be eliminated? 1623 1624 The PLD project was based on the principle that computer-searching of 1625 texts is most effective when it is done with a large database. Because 1626 PLD represented a collection that serves so many disciplines across so 1627 many periods, it was irresistible. 1628 1629 The basic rule in converting PLD was to do no harm, to avoid the sins of 1630 intrusion in such a database: no introduction of newer editions, no 1631 on-the-spot changes, no eradicating of all possible falsehoods from an 1632 edition. Thus, PLD is not the final act in electronic publishing for 1633 this discipline, but simply the beginning. The conversion of PLD has 1634 evoked numerous unanticipated questions: How will information be used? 1635 What about networking? Can the rights of a database be protected? 1636 Should one protect the rights of a database? How can it be made 1637 available? 1638 1639 Those converting PLD also tried to avoid the sins of omission, that is, 1640 excluding portions of the collections or whole sections. What about the 1641 images? PLD is full of images, some are extremely pious 1642 nineteenth-century representations of the Fathers, while others contain 1643 highly interesting elements. The goal was to cover all the text of Migne 1644 (including notes, in Greek and in Hebrew, the latter of which, in 1645 particular, causes problems in creating a search structure), all the 1646 indices, and even the images, which are being scanned in separately 1647 searchable files. 1648 1649 Several North American institutions that have placed acquisition requests 1650 for the PLD database have requested it in magnetic form without software, 1651 which means they are already running it without software, without 1652 anything demonstrated at the Workshop. 1653 1654 What cannot practically be done is go back and reconvert and re-encode 1655 data, a time-consuming and extremely costly enterprise. CALALUCA sees 1656 PLD as a database that can, and should, be run under a variety of 1657 retrieval softwares. This will permit the widest possible searches. 1658 Consequently, the need to produce a CD-ROM of PLD, as well as to develop 1659 software that could handle some 1.3 gigabyte of heavily encoded text, 1660 developed out of conversations with collection development and reference 1661 librarians who wanted software both compassionate enough for the 1662 pedestrian but also capable of incorporating the most detailed 1663 lexicographical studies that a user desires to conduct. In the end, the 1664 encoding and conversion of the data will prove the most enduring 1665 testament to the value of the project. 1666 1667 The encoding of the database was also a hard-fought issue: Did the 1668 database need to be encoded? Were there normative structures for encoding 1669 humanist texts? Should it be SGML? What about the TEI--will it last, 1670 will it prove useful? CALALUCA expressed some minor doubts as to whether 1671 a data bank can be fully TEI-conformant. Every effort can be made, but 1672 in the end to be TEI-conformant means to accept the need to make some 1673 firm encoding decisions that can, indeed, be disputed. The TEI points 1674 the publisher in a proper direction but does not presume to make all the 1675 decisions for him or her. Essentially, the goal of encoding was to 1676 eliminate, as much as possible, the hindrances to information-networking, 1677 so that if an institution acquires a database, everybody associated with 1678 the institution can have access to it. 1679 1680 CALALUCA demonstrated a portion of Volume 160, because it had the most 1681 anomalies in it. The software was created by Electronic Book 1682 Technologies of Providence, RI, and is called Dynatext. The software 1683 works only with SGML-coded data. 1684 1685 Viewing a table of contents on the screen, the audience saw how Dynatext 1686 treats each element as a book and attempts to simplify movement through a 1687 volume. Familiarity with the Patrologia in print (i.e., the text, its 1688 source, and the editions) will make the machine-readable versions highly 1689 useful. (Software with a Windows application was sought for PLD, 1690 CALALUCA said, because this was the main trend for scholarly use.) 1691 1692 CALALUCA also demonstrated how a user can perform a variety of searches 1693 and quickly move to any part of a volume; the look-up screen provides 1694 some basic, simple word-searching. 1695 1696 CALALUCA argued that one of the major difficulties is not the software. 1697 Rather, in creating a product that will be used by scholars representing 1698 a broad spectrum of computer sophistication, user documentation proves 1699 to be the most important service one can provide. 1700 1701 CALALUCA next illustrated a truncated search under mysterium within ten 1702 words of virtus and how one would be able to find its contents throughout 1703 the entire database. He said that the exciting thing about PLD is that 1704 many of the applications in the retrieval software being written for it 1705 will exceed the capabilities of the software employed now for the CD-ROM 1706 version. The CD-ROM faces genuine limitations, in terms of speed and 1707 comprehensiveness, in the creation of a retrieval software to run it. 1708 CALALUCA said he hoped that individual scholars will download the data, 1709 if they wish, to their personal computers, and have ready access to 1710 important texts on a constant basis, which they will be able to use in 1711 their research and from which they might even be able to publish. 1712 1713 (CALALUCA explained that the blue numbers represented Migne's column numbers, 1714 which are the standard scholarly references. Pulling up a note, he stated 1715 that these texts were heavily edited and the image files would appear simply 1716 as a note as well, so that one could quickly access an image.) 1717 1718 ****** 1719 1720 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1721 FLEISCHHAUER/ERWAY * Several problems with which AM is still wrestling * 1722 Various search and retrieval capabilities * Illustration of automatic 1723 stemming and a truncated search * AM's attempt to find ways to connect 1724 cataloging to the texts * AM's gravitation towards SGML * Striking a 1725 balance between quantity and quality * How AM furnishes users recourse to 1726 images * Conducting a search in a full-text environment * Macintosh and 1727 IBM prototypes of AM * Multimedia aspects of AM * 1728 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1729 1730 A demonstration of American Memory by its coordinator, Carl FLEISCHHAUER, 1731 and Ricky ERWAY, associate coordinator, Library of Congress, concluded 1732 the morning session. Beginning with a collection of broadsides from the 1733 Continental Congress and the Constitutional Convention, the only text 1734 collection in a presentable form at the time of the Workshop, FLEISCHHAUER 1735 highlighted several of the problems with which AM is still wrestling. 1736 (In its final form, the disk will contain two collections, not only the 1737 broadsides but also the full text with illustrations of a set of 1738 approximately 300 African-American pamphlets from the period 1870 to 1910.) 1739 1740 As FREEMAN had explained earlier, AM has attempted to use a small amount 1741 of interpretation to introduce collections. In the present case, the 1742 contractor, a company named Quick Source, in Silver Spring, MD., used 1743 software called Toolbook and put together a modestly interactive 1744 introduction to the collection. Like the two preceding speakers, 1745 FLEISCHHAUER argued that the real asset was the underlying collection. 1746 1747 FLEISCHHAUER proceeded to describe various search and retrieval 1748 capabilities while ERWAY worked the computer. In this particular package 1749 the "go to" pull-down allowed the user in effect to jump out of Toolbook, 1750 where the interactive program was located, and enter the third-party 1751 software used by AM for this text collection, which is called Personal 1752 Librarian. This was the Windows version of Personal Librarian, a 1753 software application put together by a company in Rockville, Md. 1754 1755 Since the broadsides came from the Revolutionary War period, a search was 1756 conducted using the words British or war, with the default operator reset 1757 as or. FLEISCHHAUER demonstrated both automatic stemming (which finds 1758 other forms of the same root) and a truncated search. One of Personal 1759 Librarian's strongest features, the relevance ranking, was represented by 1760 a chart that indicated how often words being sought appeared in 1761 documents, with the one receiving the most "hits" obtaining the highest 1762 score. The "hit list" that is supplied takes the relevance ranking into 1763 account, making the first hit, in effect, the one the software has 1764 selected as the most relevant example. 1765 1766 While in the text of one of the broadside documents, FLEISCHHAUER 1767 remarked AM's attempt to find ways to connect cataloging to the texts, 1768 which it does in different ways in different manifestations. In the case 1769 shown, the cataloging was pasted on: AM took MARC records that were 1770 written as on-line records right into one of the Library's mainframe 1771 retrieval programs, pulled them out, and handed them off to the contractor, 1772 who massaged them somewhat to display them in the manner shown. One of 1773 AM's questions is, Does the cataloguing normally performed in the mainframe 1774 work in this context, or had AM ought to think through adjustments? 1775 1776 FLEISCHHAUER made the additional point that, as far as the text goes, AM 1777 has gravitated towards SGML (he pointed to the boldface in the upper part 1778 of the screen). Although extremely limited in its ability to translate 1779 or interpret SGML, Personal Librarian will furnish both bold and italics 1780 on screen; a fairly easy thing to do, but it is one of the ways in which 1781 SGML is useful. 1782 1783 Striking a balance between quantity and quality has been a major concern 1784 of AM, with accuracy being one of the places where project staff have 1785 felt that less than 100-percent accuracy was not unacceptable. 1786 FLEISCHHAUER cited the example of the standard of the rekeying industry, 1787 namely 99.95 percent; as one service bureau informed him, to go from 1788 99.95 to 100 percent would double the cost. 1789 1790 FLEISCHHAUER next demonstrated how AM furnishes users recourse to images, 1791 and at the same time recalled LESK's pointed question concerning the 1792 number of people who would look at those images and the number who would 1793 work only with the text. If the implication of LESK's question was 1794 sound, FLEISCHHAUER said, it raised the stakes for text accuracy and 1795 reduced the value of the strategy for images. 1796 1797 Contending that preservation is always a bugaboo, FLEISCHHAUER 1798 demonstrated several images derived from a scan of a preservation 1799 microfilm that AM had made. He awarded a grade of C at best, perhaps a 1800 C minus or a C plus, for how well it worked out. Indeed, the matter of 1801 learning if other people had better ideas about scanning in general, and, 1802 in particular, scanning from microfilm, was one of the factors that drove 1803 AM to attempt to think through the agenda for the Workshop. Skew, for 1804 example, was one of the issues that AM in its ignorance had not reckoned 1805 would prove so difficult. 1806 1807 Further, the handling of images of the sort shown, in a desktop computer 1808 environment, involved a considerable amount of zooming and scrolling. 1809 Ultimately, AM staff feel that perhaps the paper copy that is printed out 1810 might be the most useful one, but they remain uncertain as to how much 1811 on-screen reading users will do. 1812 1813 Returning to the text, FLEISCHHAUER asked viewers to imagine a person who 1814 might be conducting a search in a full-text environment. With this 1815 scenario, he proceeded to illustrate other features of Personal Librarian 1816 that he considered helpful; for example, it provides the ability to 1817 notice words as one reads. Clicking the "include" button on the bottom 1818 of the search window pops the words that have been highlighted into the 1819 search. Thus, a user can refine the search as he or she reads, 1820 re-executing the search and continuing to find things in the quest for 1821 materials. This software not only contains relevance ranking, Boolean 1822 operators, and truncation, it also permits one to perform word algebra, 1823 so to say, where one puts two or three words in parentheses and links 1824 them with one Boolean operator and then a couple of words in another set 1825 of parentheses and asks for things within so many words of others. 1826 1827 Until they became acquainted recently with some of the work being done in 1828 classics, the AM staff had not realized that a large number of the 1829 projects that involve electronic texts were being done by people with a 1830 profound interest in language and linguistics. Their search strategies 1831 and thinking are oriented to those fields, as is shown in particular by 1832 the Perseus example. As amateur historians, the AM staff were thinking 1833 more of searching for concepts and ideas than for particular words. 1834 Obviously, FLEISCHHAUER conceded, searching for concepts and ideas and 1835 searching for words may be two rather closely related things. 1836 1837 While displaying several images, FLEISCHHAUER observed that the Macintosh 1838 prototype built by AM contains a greater diversity of formats. Echoing a 1839 previous speaker, he said that it was easier to stitch things together in 1840 the Macintosh, though it tended to be a little more anemic in search and 1841 retrieval. AM, therefore, increasingly has been investigating 1842 sophisticated retrieval engines in the IBM format. 1843 1844 FLEISCHHAUER demonstrated several additional examples of the prototype 1845 interfaces: One was AM's metaphor for the network future, in which a 1846 kind of reading-room graphic suggests how one would be able to go around 1847 to different materials. AM contains a large number of photographs in 1848 analog video form worked up from a videodisc, which enable users to make 1849 copies to print or incorporate in digital documents. A frame-grabber is 1850 built into the system, making it possible to bring an image into a window 1851 and digitize or print it out. 1852 1853 FLEISCHHAUER next demonstrated sound recording, which included texts. 1854 Recycled from a previous project, the collection included sixty 78-rpm 1855 phonograph records of political speeches that were made during and 1856 immediately after World War I. These constituted approximately three 1857 hours of audio, as AM has digitized it, which occupy 150 megabytes on a 1858 CD. Thus, they are considerably compressed. From the catalogue card, 1859 FLEISCHHAUER proceeded to a transcript of a speech with the audio 1860 available and with highlighted text following it as it played. 1861 A photograph has been added and a transcription made. 1862 1863 Considerable value has been added beyond what the Library of Congress 1864 normally would do in cataloguing a sound recording, which raises several 1865 questions for AM concerning where to draw lines about how much value it can 1866 afford to add and at what point, perhaps, this becomes more than AM could 1867 reasonably do or reasonably wish to do. FLEISCHHAUER also demonstrated 1868 a motion picture. As FREEMAN had reported earlier, the motion picture 1869 materials have proved the most popular, not surprisingly. This says more 1870 about the medium, he thought, than about AM's presentation of it. 1871 1872 Because AM's goal was to bring together things that could be used by 1873 historians or by people who were curious about history, 1874 turn-of-the-century footage seemed to represent the most appropriate 1875 collections from the Library of Congress in motion pictures. These were 1876 the very first films made by Thomas Edison's company and some others at 1877 that time. The particular example illustrated was a Biograph film, 1878 brought in with a frame-grabber into a window. A single videodisc 1879 contains about fifty titles and pieces of film from that period, all of 1880 New York City. Taken together, AM believes, they provide an interesting 1881 documentary resource. 1882 1883 ****** 1884 1885 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1886 DISCUSSION * Using the frame-grabber in AM * Volume of material processed 1887 and to be processed * Purpose of AM within LC * Cataloguing and the 1888 nature of AM's material * SGML coding and the question of quality versus 1889 quantity * 1890 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1891 1892 During the question-and-answer period that followed FLEISCHHAUER's 1893 presentation, several clarifications were made. 1894 1895 AM is bringing in motion pictures from a videodisc. The frame-grabber 1896 devices create a window on a computer screen, which permits users to 1897 digitize a single frame of the movie or one of the photographs. It 1898 produces a crude, rough-and-ready image that high school students can 1899 incorporate into papers, and that has worked very nicely in this way. 1900 1901 Commenting on FLEISCHHAUER's assertion that AM was looking more at 1902 searching ideas than words, MYLONAS argued that without words an idea 1903 does not exist. FLEISCHHAUER conceded that he ought to have articulated 1904 his point more clearly. MYLONAS stated that they were in fact both 1905 talking about the same thing. By searching for words and by forcing 1906 people to focus on the word, the Perseus Project felt that they would get 1907 them to the idea. The way one reviews results is tailored more to one 1908 kind of user than another. 1909 1910 Concerning the total volume of material that has been processed in this 1911 way, AM at this point has in retrievable form seven or eight collections, 1912 all of them photographic. In the Macintosh environment, for example, 1913 there probably are 35,000-40,000 photographs. The sound recordings 1914 number sixty items. The broadsides number about 300 items. There are 1915 500 political cartoons in the form of drawings. The motion pictures, as 1916 individual items, number sixty to seventy. 1917 1918 AM also has a manuscript collection, the life history portion of one of 1919 the federal project series, which will contain 2,900 individual 1920 documents, all first-person narratives. AM has in process about 350 1921 African-American pamphlets, or about 12,000 printed pages for the period 1922 1870-1910. Also in the works are some 4,000 panoramic photographs. AM 1923 has recycled a fair amount of the work done by LC's Prints and 1924 Photographs Division during the Library's optical disk pilot project in 1925 the 1980s. For example, a special division of LC has tooled up and 1926 thought through all the ramifications of electronic presentation of 1927 photographs. Indeed, they are wheeling them out in great barrel loads. 1928 The purpose of AM within the Library, it is hoped, is to catalyze several 1929 of the other special collection divisions which have no particular 1930 experience with, in some cases, mixed feelings about, an activity such as 1931 AM. Moreover, in many cases the divisions may be characterized as not 1932 only lacking experience in "electronifying" things but also in automated 1933 cataloguing. MARC cataloguing as practiced in the United States is 1934 heavily weighted toward the description of monograph and serial 1935 materials, but is much thinner when one enters the world of manuscripts 1936 and things that are held in the Library's music collection and other 1937 units. In response to a comment by LESK, that AM's material is very 1938 heavily photographic, and is so primarily because individual records have 1939 been made for each photograph, FLEISCHHAUER observed that an item-level 1940 catalog record exists, for example, for each photograph in the Detroit 1941 Publishing collection of 25,000 pictures. In the case of the Federal 1942 Writers Project, for which nearly 3,000 documents exist, representing 1943 information from twenty-six different states, AM with the assistance of 1944 Karen STUART of the Manuscript Division will attempt to find some way not 1945 only to have a collection-level record but perhaps a MARC record for each 1946 state, which will then serve as an umbrella for the 100-200 documents 1947 that come under it. But that drama remains to be enacted. The AM staff 1948 is conservative and clings to cataloguing, though of course visitors tout 1949 artificial intelligence and neural networks in a manner that suggests that 1950 perhaps one need not have cataloguing or that much of it could be put aside. 1951 1952 The matter of SGML coding, FLEISCHHAUER conceded, returned the discussion 1953 to the earlier treated question of quality versus quantity in the Library 1954 of Congress. Of course, text conversion can be done with 100-percent 1955 accuracy, but it means that when one's holdings are as vast as LC's only 1956 a tiny amount will be exposed, whereas permitting lower levels of 1957 accuracy can lead to exposing or sharing larger amounts, but with the 1958 quality correspondingly impaired. 1959 1960 ****** 1961 1962 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1963 TWOHIG * A contrary experience concerning electronic options * Volume of 1964 material in the Washington papers and a suggestion of David Packard * 1965 Implications of Packard's suggestion * Transcribing the documents for the 1966 CD-ROM * Accuracy of transcriptions * The CD-ROM edition of the Founding 1967 Fathers documents * 1968 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1969 1970 Finding encouragement in a comment of MICHELSON's from the morning 1971 session--that numerous people in the humanities were choosing electronic 1972 options to do their work--Dorothy TWOHIG, editor, The Papers of George 1973 Washington, opened her illustrated talk by noting that her experience 1974 with literary scholars and numerous people in editing was contrary to 1975 MICHELSON's. TWOHIG emphasized literary scholars' complete ignorance of 1976 the technological options available to them or their reluctance or, in 1977 some cases, their downright hostility toward these options. 1978 1979 After providing an overview of the five Founding Fathers projects 1980 (Jefferson at Princeton, Franklin at Yale, John Adams at the 1981 Massachusetts Historical Society, and Madison down the hall from her at 1982 the University of Virginia), TWOHIG observed that the Washington papers, 1983 like all of the projects, include both sides of the Washington 1984 correspondence and deal with some 135,000 documents to be published with 1985 extensive annotation in eighty to eighty-five volumes, a project that 1986 will not be completed until well into the next century. Thus, it was 1987 with considerable enthusiasm several years ago that the Washington Papers 1988 Project (WPP) greeted David Packard's suggestion that the papers of the 1989 Founding Fathers could be published easily and inexpensively, and to the 1990 great benefit of American scholarship, via CD-ROM. 1991 1992 In pragmatic terms, funding from the Packard Foundation would expedite 1993 the transcription of thousands of documents waiting to be put on disk in 1994 the WPP offices. Further, since the costs of collecting, editing, and 1995 converting the Founding Fathers documents into letterpress editions were 1996 running into the millions of dollars, and the considerable staffs 1997 involved in all of these projects were devoting their careers to 1998 producing the work, the Packard Foundation's suggestion had a 1999 revolutionary aspect: Transcriptions of the entire corpus of the 2000 Founding Fathers papers would be available on CD-ROM to public and 2001 college libraries, even high schools, at a fraction of the cost-- 2002 $100-$150 for the annual license fee--to produce a limited university 2003 press run of 1,000 of each volume of the published papers at $45-$150 per 2004 printed volume. Given the current budget crunch in educational systems 2005 and the corresponding constraints on librarians in smaller institutions 2006 who wish to add these volumes to their collections, producing the 2007 documents on CD-ROM would likely open a greatly expanded audience for the 2008 papers. TWOHIG stressed, however, that development of the Founding 2009 Fathers CD-ROM is still in its infancy. Serious software problems remain 2010 to be resolved before the material can be put into readable form. 2011 2012 Funding from the Packard Foundation resulted in a major push to 2013 transcribe the 75,000 or so documents of the Washington papers remaining 2014 to be transcribed onto computer disks. Slides illustrated several of the 2015 problems encountered, for example, the present inability of CD-ROM to 2016 indicate the cross-outs (deleted material) in eighteenth century 2017 documents. TWOHIG next described documents from various periods in the 2018 eighteenth century that have been transcribed in chronological order and 2019 delivered to the Packard offices in California, where they are converted 2020 to the CD-ROM, a process that is expected to consume five years to 2021 complete (that is, reckoning from David Packard's suggestion made several 2022 years ago, until about July 1994). TWOHIG found an encouraging 2023 indication of the project's benefits in the ongoing use made by scholars 2024 of the search functions of the CD-ROM, particularly in reducing the time 2025 spent in manually turning the pages of the Washington papers. 2026 2027 TWOHIG next furnished details concerning the accuracy of transcriptions. 2028 For instance, the insertion of thousands of documents on the CD-ROM 2029 currently does not permit each document to be verified against the 2030 original manuscript several times as in the case of documents that appear 2031 in the published edition. However, the transcriptions receive a cursory 2032 check for obvious typos, the misspellings of proper names, and other 2033 errors from the WPP CD-ROM editor. Eventually, all documents that appear 2034 in the electronic version will be checked by project editors. Although 2035 this process has met with opposition from some of the editors on the 2036 grounds that imperfect work may leave their offices, the advantages in 2037 making this material available as a research tool outweigh fears about the 2038 misspelling of proper names and other relatively minor editorial matters. 2039 2040 Completion of all five Founding Fathers projects (i.e., retrievability 2041 and searchability of all of the documents by proper names, alternate 2042 spellings, or varieties of subjects) will provide one of the richest 2043 sources of this size for the history of the United States in the latter 2044 part of the eighteenth century. Further, publication on CD-ROM will 2045 allow editors to include even minutiae, such as laundry lists, not 2046 included in the printed volumes. 2047 2048 It seems possible that the extensive annotation provided in the printed 2049 volumes eventually will be added to the CD-ROM edition, pending 2050 negotiations with the publishers of the papers. At the moment, the 2051 Founding Fathers CD-ROM is accessible only on the IBYCUS, a computer 2052 developed out of the Thesaurus Linguae Graecae project and designed for 2053 the use of classical scholars. There are perhaps 400 IBYCUS computers in 2054 the country, most of which are in university classics departments. 2055 Ultimately, it is anticipated that the CD-ROM edition of the Founding 2056 Fathers documents will run on any IBM-compatible or Macintosh computer 2057 with a CD-ROM drive. Numerous changes in the software will also occur 2058 before the project is completed. (Editor's note: an IBYCUS was 2059 unavailable to demonstrate the CD-ROM.) 2060 2061 ****** 2062 2063 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 2064 DISCUSSION * Several additional features of WPP clarified * 2065 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 2066 2067 Discussion following TWOHIG's presentation served to clarify several 2068 additional features, including (1) that the project's primary 2069 intellectual product consists in the electronic transcription of the 2070 material; (2) that the text transmitted to the CD-ROM people is not 2071 marked up; (3) that cataloging and subject-indexing of the material 2072 remain to be worked out (though at this point material can be retrieved 2073 by name); and (4) that because all the searching is done in the hardware, 2074 the IBYCUS is designed to read a CD-ROM which contains only sequential 2075 text files. Technically, it then becomes very easy to read the material 2076 off and put it on another device. 2077 2078 ****** 2079 2080 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 2081 LEBRON * Overview of the history of the joint project between AAAS and 2082 OCLC * Several practices the on-line environment shares with traditional 2083 publishing on hard copy * Several technical and behavioral barriers to 2084 electronic publishing * How AAAS and OCLC arrived at the subject of 2085 clinical trials * Advantages of the electronic format and other features 2086 of OJCCT * An illustrated tour of the journal * 2087 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 2088 2089 Maria LEBRON, managing editor, The Online Journal of Current Clinical 2090 Trials (OJCCT), presented an illustrated overview of the history of the 2091 joint project between the American Association for the Advancement of 2092 Science (AAAS) and the Online Computer Library Center, Inc. (OCLC). The 2093 joint venture between AAAS and OCLC owes its beginning to a 2094 reorganization launched by the new chief executive officer at OCLC about 2095 three years ago and combines the strengths of these two disparate 2096 organizations. In short, OJCCT represents the process of scholarly 2097 publishing on line. 2098 2099 LEBRON next discussed several practices the on-line environment shares 2100 with traditional publishing on hard copy--for example, peer review of 2101 manuscripts--that are highly important in the academic world. LEBRON 2102 noted in particular the implications of citation counts for tenure 2103 committees and grants committees. In the traditional hard-copy 2104 environment, citation counts are readily demonstrable, whereas the 2105 on-line environment represents an ethereal medium to most academics. 2106 2107 LEBRON remarked several technical and behavioral barriers to electronic 2108 publishing, for instance, the problems in transmission created by special 2109 characters or by complex graphics and halftones. In addition, she noted 2110 economic limitations such as the storage costs of maintaining back issues 2111 and market or audience education. 2112 2113 Manuscripts cannot be uploaded to OJCCT, LEBRON explained, because it is 2114 not a bulletin board or E-mail, forms of electronic transmission of 2115 information that have created an ambience clouding people's understanding 2116 of what the journal is attempting to do. OJCCT, which publishes 2117 peer-reviewed medical articles dealing with the subject of clinical 2118 trials, includes text, tabular material, and graphics, although at this 2119 time it can transmit only line illustrations. 2120 2121 Next, LEBRON described how AAAS and OCLC arrived at the subject of 2122 clinical trials: It is 1) a highly statistical discipline that 2) does 2123 not require halftones but can satisfy the needs of its audience with line 2124 illustrations and graphic material, and 3) there is a need for the speedy 2125 dissemination of high-quality research results. Clinical trials are 2126 research activities that involve the administration of a test treatment 2127 to some experimental unit in order to test its usefulness before it is 2128 made available to the general population. LEBRON proceeded to give 2129 additional information on OJCCT concerning its editor-in-chief, editorial 2130 board, editorial content, and the types of articles it publishes 2131 (including peer-reviewed research reports and reviews), as well as 2132 features shared by other traditional hard-copy journals. 2133 2134 Among the advantages of the electronic format are faster dissemination of 2135 information, including raw data, and the absence of space constraints 2136 because pages do not exist. (This latter fact creates an interesting 2137 situation when it comes to citations.) Nor are there any issues. AAAS's 2138 capacity to download materials directly from the journal to a 2139 subscriber's printer, hard drive, or floppy disk helps ensure highly 2140 accurate transcription. Other features of OJCCT include on-screen alerts 2141 that allow linkage of subsequently published documents to the original 2142 documents; on-line searching by subject, author, title, etc.; indexing of 2143 every single word that appears in an article; viewing access to an 2144 article by component (abstract, full text, or graphs); numbered 2145 paragraphs to replace page counts; publication in Science every thirty 2146 days of indexing of all articles published in the journal; 2147 typeset-quality screens; and Hypertext links that enable subscribers to 2148 bring up Medline abstracts directly without leaving the journal. 2149 2150 After detailing the two primary ways to gain access to the journal, 2151 through the OCLC network and Compuserv if one desires graphics or through 2152 the Internet if just an ASCII file is desired, LEBRON illustrated the 2153 speedy editorial process and the coding of the document using SGML tags 2154 after it has been accepted for publication. She also gave an illustrated 2155 tour of the journal, its search-and-retrieval capabilities in particular, 2156 but also including problems associated with scanning in illustrations, 2157 and the importance of on-screen alerts to the medical profession re 2158 retractions or corrections, or more frequently, editorials, letters to 2159 the editors, or follow-up reports. She closed by inviting the audience 2160 to join AAAS on 1 July, when OJCCT was scheduled to go on-line. 2161 2162 ****** 2163 2164 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 2165 DISCUSSION * Additional features of OJCCT * 2166 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 2167 2168 In the lengthy discussion that followed LEBRON's presentation, these 2169 points emerged: 2170 2171 * The SGML text can be tailored as users wish. 2172 2173 * All these articles have a fairly simple document definition. 2174 2175 * Document-type definitions (DTDs) were developed and given to OJCCT 2176 for coding. 2177 2178 * No articles will be removed from the journal. (Because there are 2179 no back issues, there are no lost issues either. Once a subscriber 2180 logs onto the journal he or she has access not only to the currently 2181 published materials, but retrospectively to everything that has been 2182 published in it. Thus the table of contents grows bigger. The date 2183 of publication serves to distinguish between currently published 2184 materials and older materials.) 2185 2186 * The pricing system for the journal resembles that for most medical 2187 journals: for 1992, $95 for a year, plus telecommunications charges 2188 (there are no connect time charges); for 1993, $110 for the 2189 entire year for single users, though the journal can be put on a 2190 local area network (LAN). However, only one person can access the 2191 journal at a time. Site licenses may come in the future. 2192 2193 * AAAS is working closely with colleagues at OCLC to display 2194 mathematical equations on screen. 2195 2196 * Without compromising any steps in the editorial process, the 2197 technology has reduced the time lag between when a manuscript is 2198 originally submitted and the time it is accepted; the review process 2199 does not differ greatly from the standard six-to-eight weeks 2200 employed by many of the hard-copy journals. The process still 2201 depends on people. 2202 2203 * As far as a preservation copy is concerned, articles will be 2204 maintained on the computer permanently and subscribers, as part of 2205 their subscription, will receive a microfiche-quality archival copy 2206 of everything published during that year; in addition, reprints can 2207 be purchased in much the same way as in a hard-copy environment. 2208 Hard copies are prepared but are not the primary medium for the 2209 dissemination of the information. 2210 2211 * Because OJCCT is not yet on line, it is difficult to know how many 2212 people would simply browse through the journal on the screen as 2213 opposed to downloading the whole thing and printing it out; a mix of 2214 both types of users likely will result. 2215 2216 ****** 2217 2218 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 2219 PERSONIUS * Developments in technology over the past decade * The CLASS 2220 Project * Advantages for technology and for the CLASS Project * 2221 Developing a network application an underlying assumption of the project 2222 * Details of the scanning process * Print-on-demand copies of books * 2223 Future plans include development of a browsing tool * 2224 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 2225 2226 Lynne PERSONIUS, assistant director, Cornell Information Technologies for 2227 Scholarly Information Services, Cornell University, first commented on 2228 the tremendous impact that developments in technology over the past ten 2229 years--networking, in particular--have had on the way information is 2230 handled, and how, in her own case, these developments have counterbalanced 2231 Cornell's relative geographical isolation. Other significant technologies 2232 include scanners, which are much more sophisticated than they were ten years 2233 ago; mass storage and the dramatic savings that result from it in terms of 2234 both space and money relative to twenty or thirty years ago; new and 2235 improved printing technologies, which have greatly affected the distribution 2236 of information; and, of course, digital technologies, whose applicability to 2237 library preservation remains at issue. 2238 2239 Given that context, PERSONIUS described the College Library Access and 2240 Storage System (CLASS) Project, a library preservation project, 2241 primarily, and what has been accomplished. Directly funded by the 2242 Commission on Preservation and Access and by the Xerox Corporation, which 2243 has provided a significant amount of hardware, the CLASS Project has been 2244 working with a development team at Xerox to develop a software 2245 application tailored to library preservation requirements. Within 2246 Cornell, participants in the project have been working jointly with both 2247 library and information technologies. The focus of the project has been 2248 on reformatting and saving books that are in brittle condition. 2249 PERSONIUS showed Workshop participants a brittle book, and described how 2250 such books were the result of developments in papermaking around the 2251 beginning of the Industrial Revolution. The papermaking process was 2252 changed so that a significant amount of acid was introduced into the 2253 actual paper itself, which deteriorates as it sits on library shelves. 2254 2255 One of the advantages for technology and for the CLASS Project is that 2256 the information in brittle books is mostly out of copyright and thus 2257 offers an opportunity to work with material that requires library 2258 preservation, and to create and work on an infrastructure to save the 2259 material. Acknowledging the familiarity of those working in preservation 2260 with this information, PERSONIUS noted that several things are being 2261 done: the primary preservation technology used today is photocopying of 2262 brittle material. Saving the intellectual content of the material is the 2263 main goal. With microfilm copy, the intellectual content is preserved on 2264 the assumption that in the future the image can be reformatted in any 2265 other way that then exists. 2266 2267 An underlying assumption of the CLASS Project from the beginning was 2268 that it would develop a network application. Project staff scan books 2269 at a workstation located in the library, near the brittle material. 2270 An image-server filing system is located at a distance from that 2271 workstation, and a printer is located in another building. All of the 2272 materials digitized and stored on the image-filing system are cataloged 2273 in the on-line catalogue. In fact, a record for each of these electronic 2274 books is stored in the RLIN database so that a record exists of what is 2275 in the digital library throughout standard catalogue procedures. In the 2276 future, researchers working from their own workstations in their offices, 2277 or their networks, will have access--wherever they might be--through a 2278 request server being built into the new digital library. A second 2279 assumption is that the preferred means of finding the material will be by 2280 looking through a catalogue. PERSONIUS described the scanning process, 2281 which uses a prototype scanner being developed by Xerox and which scans a 2282 very high resolution image at great speed. Another significant feature, 2283 because this is a preservation application, is the placing of the pages 2284 that fall apart one for one on the platen. Ordinarily, a scanner could 2285 be used with some sort of a document feeder, but because of this 2286 application that is not feasible. Further, because CLASS is a 2287 preservation application, after the paper replacement is made there, a 2288 very careful quality control check is performed. An original book is 2289 compared to the printed copy and verification is made, before proceeding, 2290 that all of the image, all of the information, has been captured. Then, 2291 a new library book is produced: The printed images are rebound by a 2292 commercial binder and a new book is returned to the shelf. 2293 Significantly, the books returned to the library shelves are beautiful 2294 and useful replacements on acid-free paper that should last a long time, 2295 in effect, the equivalent of preservation photocopies. Thus, the project 2296 has a library of digital books. In essence, CLASS is scanning and 2297 storing books as 600 dot-per-inch bit-mapped images, compressed using 2298 Group 4 CCITT (i.e., the French acronym for International Consultative 2299 Committee for Telegraph and Telephone) compression. They are stored as 2300 TIFF files on an optical filing system that is composed of a database 2301 used for searching and locating the books and an optical jukebox that 2302 stores 64 twelve-inch platters. A very-high-resolution printed copy of 2303 these books at 600 dots per inch is created, using a Xerox DocuTech 2304 printer to make the paper replacements on acid-free paper. 2305 2306 PERSONIUS maintained that the CLASS Project presents an opportunity to 2307 introduce people to books as digital images by using a paper medium. 2308 Books are returned to the shelves while people are also given the ability 2309 to print on demand--to make their own copies of books. (PERSONIUS 2310 distributed copies of an engineering journal published by engineering 2311 students at Cornell around 1900 as an example of what a print-on-demand 2312 copy of material might be like. This very cheap copy would be available 2313 to people to use for their own research purposes and would bridge the gap 2314 between an electronic work and the paper that readers like to have.) 2315 PERSONIUS then attempted to illustrate a very early prototype of 2316 networked access to this digital library. Xerox Corporation has 2317 developed a prototype of a view station that can send images across the 2318 network to be viewed. 2319 2320 The particular library brought down for demonstration contained two 2321 mathematics books. CLASS is developing and will spend the next year 2322 developing an application that allows people at workstations to browse 2323 the books. Thus, CLASS is developing a browsing tool, on the assumption 2324 that users do not want to read an entire book from a workstation, but 2325 would prefer to be able to look through and decide if they would like to 2326 have a printed copy of it. 2327 2328 ****** 2329 2330 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 2331 DISCUSSION * Re retrieval software * "Digital file copyright" * Scanning 2332 rate during production * Autosegmentation * Criteria employed in 2333 selecting books for scanning * Compression and decompression of images * 2334 OCR not precluded * 2335 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 2336 2337 During the question-and-answer period that followed her presentation, 2338 PERSONIUS made these additional points: 2339 2340 * Re retrieval software, Cornell is developing a Unix-based server 2341 as well as clients for the server that support multiple platforms 2342 (Macintosh, IBM and Sun workstations), in the hope that people from 2343 any of those platforms will retrieve books; a further operating 2344 assumption is that standard interfaces will be used as much as 2345 possible, where standards can be put in place, because CLASS 2346 considers this retrieval software a library application and would 2347 like to be able to look at material not only at Cornell but at other 2348 institutions. 2349 2350 * The phrase "digital file copyright by Cornell University" was 2351 added at the advice of Cornell's legal staff with the caveat that it 2352 probably would not hold up in court. Cornell does not want people 2353 to copy its books and sell them but would like to keep them 2354 available for use in a library environment for library purposes. 2355 2356 * In production the scanner can scan about 300 pages per hour, 2357 capturing 600 dots per inch. 2358 2359 * The Xerox software has filters to scan halftone material and avoid 2360 the moire patterns that occur when halftone material is scanned. 2361 Xerox has been working on hardware and software that would enable 2362 the scanner itself to recognize this situation and deal with it 2363 appropriately--a kind of autosegmentation that would enable the 2364 scanner to handle halftone material as well as text on a single page. 2365 2366 * The books subjected to the elaborate process described above were 2367 selected because CLASS is a preservation project, with the first 500 2368 books selected coming from Cornell's mathematics collection, because 2369 they were still being heavily used and because, although they were 2370 in need of preservation, the mathematics library and the mathematics 2371 faculty were uncomfortable having them microfilmed. (They wanted a 2372 printed copy.) Thus, these books became a logical choice for this 2373 project. Other books were chosen by the project's selection committees 2374 for experiments with the technology, as well as to meet a demand or need. 2375 2376 * Images will be decompressed before they are sent over the line; at 2377 this time they are compressed and sent to the image filing system 2378 and then sent to the printer as compressed images; they are returned 2379 to the workstation as compressed 600-dpi images and the workstation 2380 decompresses and scales them for display--an inefficient way to 2381 access the material though it works quite well for printing and 2382 other purposes. 2383 2384 * CLASS is also decompressing on Macintosh and IBM, a slow process 2385 right now. Eventually, compression and decompression will take 2386 place on an image conversion server. Trade-offs will be made, based 2387 on future performance testing, concerning where the file is 2388 compressed and what resolution image is sent. 2389 2390 * OCR has not been precluded; images are being stored that have been 2391 scanned at a high resolution, which presumably would suit them well 2392 to an OCR process. Because the material being scanned is about 100 2393 years old and was printed with less-than-ideal technologies, very 2394 early and preliminary tests have not produced good results. But the 2395 project is capturing an image that is of sufficient resolution to be 2396 subjected to OCR in the future. Moreover, the system architecture 2397 and the system plan have a logical place to store an OCR image if it 2398 has been captured. But that is not being done now. 2399 2400 ****** 2401 2402 SESSION III. DISTRIBUTION, NETWORKS, AND NETWORKING: OPTIONS FOR 2403 DISSEMINATION 2404 2405 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 2406 ZICH * Issues pertaining to CD-ROMs * Options for publishing in CD-ROM * 2407 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 2408 2409 Robert ZICH, special assistant to the associate librarian for special 2410 projects, Library of Congress, and moderator of this session, first noted 2411 the blessed but somewhat awkward circumstance of having four very 2412 distinguished people representing networks and networking or at least 2413 leaning in that direction, while lacking anyone to speak from the 2414 strongest possible background in CD-ROMs. ZICH expressed the hope that 2415 members of the audience would join the discussion. He stressed the 2416 subtitle of this particular session, "Options for Dissemination," and, 2417 concerning CD-ROMs, the importance of determining when it would be wise 2418 to consider dissemination in CD-ROM versus networks. A shopping list of 2419 issues pertaining to CD-ROMs included: the grounds for selecting 2420 commercial publishers, and in-house publication where possible versus 2421 nonprofit or government publication. A similar list for networks 2422 included: determining when one should consider dissemination through a 2423 network, identifying the mechanisms or entities that exist to place items 2424 on networks, identifying the pool of existing networks, determining how a 2425 producer would choose between networks, and identifying the elements of 2426 a business arrangement in a network. 2427 2428 Options for publishing in CD-ROM: an outside publisher versus 2429 self-publication. If an outside publisher is used, it can be nonprofit, 2430 such as the Government Printing Office (GPO) or the National Technical 2431 Information Service (NTIS), in the case of government. The pros and cons 2432 associated with employing an outside publisher are obvious. Among the 2433 pros, there is no trouble getting accepted. One pays the bill and, in 2434 effect, goes one's way. Among the cons, when one pays an outside 2435 publisher to perform the work, that publisher will perform the work it is 2436 obliged to do, but perhaps without the production expertise and skill in 2437 marketing and dissemination that some would seek. There is the body of 2438 commercial publishers that do possess that kind of expertise in 2439 distribution and marketing but that obviously are selective. In 2440 self-publication, one exercises full control, but then one must handle 2441 matters such as distribution and marketing. Such are some of the options 2442 for publishing in the case of CD-ROM. 2443 2444 In the case of technical and design issues, which are also important, 2445 there are many matters which many at the Workshop already knew a good 2446 deal about: retrieval system requirements and costs, what to do about 2447 images, the various capabilities and platforms, the trade-offs between 2448 cost and performance, concerns about local-area networkability, 2449 interoperability, etc. 2450 2451 ****** 2452 2453 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 2454 LYNCH * Creating networked information is different from using networks 2455 as an access or dissemination vehicle * Networked multimedia on a large 2456 scale does not yet work * Typical CD-ROM publication model a two-edged 2457 sword * Publishing information on a CD-ROM in the present world of 2458 immature standards * Contrast between CD-ROM and network pricing * 2459 Examples demonstrated earlier in the day as a set of insular information 2460 gems * Paramount need to link databases * Layering to become increasingly 2461 necessary * Project NEEDS and the issues of information reuse and active 2462 versus passive use * X-Windows as a way of differentiating between 2463 network access and networked information * Barriers to the distribution 2464 of networked multimedia information * Need for good, real-time delivery 2465 protocols * The question of presentation integrity in client-server 2466 computing in the academic world * Recommendations for producing multimedia 2467 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 2468 2469 Clifford LYNCH, director, Library Automation, University of California, 2470 opened his talk with the general observation that networked information 2471 constituted a difficult and elusive topic because it is something just 2472 starting to develop and not yet fully understood. LYNCH contended that 2473 creating genuinely networked information was different from using 2474 networks as an access or dissemination vehicle and was more sophisticated 2475 and more subtle. He invited the members of the audience to extrapolate, 2476 from what they heard about the preceding demonstration projects, to what 2477 sort of a world of electronics information--scholarly, archival, 2478 cultural, etc.--they wished to end up with ten or fifteen years from now. 2479 LYNCH suggested that to extrapolate directly from these projects would 2480 produce unpleasant results. 2481 2482 Putting the issue of CD-ROM in perspective before getting into 2483 generalities on networked information, LYNCH observed that those engaged 2484 in multimedia today who wish to ship a product, so to say, probably do 2485 not have much choice except to use CD-ROM: networked multimedia on a 2486 large scale basically does not yet work because the technology does not 2487 exist. For example, anybody who has tried moving images around over the 2488 Internet knows that this is an exciting touch-and-go process, a 2489 fascinating and fertile area for experimentation, research, and 2490 development, but not something that one can become deeply enthusiastic 2491 about committing to production systems at this time. 2492 2493 This situation will change, LYNCH said. He differentiated CD-ROM from 2494 the practices that have been followed up to now in distributing data on 2495 CD-ROM. For LYNCH the problem with CD-ROM is not its portability or its 2496 slowness but the two-edged sword of having the retrieval application and 2497 the user interface inextricably bound up with the data, which is the 2498 typical CD-ROM publication model. It is not a case of publishing data 2499 but of distributing a typically stand-alone, typically closed system, 2500 all--software, user interface, and data--on a little disk. Hence, all 2501 the between-disk navigational issues as well as the impossibility in most 2502 cases of integrating data on one disk with that on another. Most CD-ROM 2503 retrieval software does not network very gracefully at present. However, 2504 in the present world of immature standards and lack of understanding of 2505 what network information is or what the ground rules are for creating or 2506 using it, publishing information on a CD-ROM does add value in a very 2507 real sense. 2508 2509 LYNCH drew a contrast between CD-ROM and network pricing and in doing so 2510 highlighted something bizarre in information pricing. A large 2511 institution such as the University of California has vendors who will 2512 offer to sell information on CD-ROM for a price per year in four digits, 2513 but for the same data (e.g., an abstracting and indexing database) on 2514 magnetic tape, regardless of how many people may use it concurrently, 2515 will quote a price in six digits. 2516 2517 What is packaged with the CD-ROM in one sense adds value--a complete 2518 access system, not just raw, unrefined information--although it is not 2519 generally perceived that way. This is because the access software, 2520 although it adds value, is viewed by some people, particularly in the 2521 university environment where there is a very heavy commitment to 2522 networking, as being developed in the wrong direction. 2523 2524 Given that context, LYNCH described the examples demonstrated as a set of 2525 insular information gems--Perseus, for example, offers nicely linked 2526 information, but would be very difficult to integrate with other 2527 databases, that is, to link together seamlessly with other source files 2528 from other sources. It resembles an island, and in this respect is 2529 similar to numerous stand-alone projects that are based on videodiscs, 2530 that is, on the single-workstation concept. 2531 2532 As scholarship evolves in a network environment, the paramount need will 2533 be to link databases. We must link personal databases to public 2534 databases, to group databases, in fairly seamless ways--which is 2535 extremely difficult in the environments under discussion with copies of 2536 databases proliferating all over the place. 2537 2538 The notion of layering also struck LYNCH as lurking in several of the 2539 projects demonstrated. Several databases in a sense constitute 2540 information archives without a significant amount of navigation built in. 2541 Educators, critics, and others will want a layered structure--one that 2542 defines or links paths through the layers to allow users to reach 2543 specific points. In LYNCH's view, layering will become increasingly 2544 necessary, and not just within a single resource but across resources 2545 (e.g., tracing mythology and cultural themes across several classics 2546 databases as well as a database of Renaissance culture). This ability to 2547 organize resources, to build things out of multiple other things on the 2548 network or select pieces of it, represented for LYNCH one of the key 2549 aspects of network information. 2550 2551 Contending that information reuse constituted another significant issue, 2552 LYNCH commended to the audience's attention Project NEEDS (i.e., National 2553 Engineering Education Delivery System). This project's objective is to 2554 produce a database of engineering courseware as well as the components 2555 that can be used to develop new courseware. In a number of the existing 2556 applications, LYNCH said, the issue of reuse (how much one can take apart 2557 and reuse in other applications) was not being well considered. He also 2558 raised the issue of active versus passive use, one aspect of which is 2559 how much information will be manipulated locally by users. Most people, 2560 he argued, may do a little browsing and then will wish to print. LYNCH 2561 was uncertain how these resources would be used by the vast majority of 2562 users in the network environment. 2563 2564 LYNCH next said a few words about X-Windows as a way of differentiating 2565 between network access and networked information. A number of the 2566 applications demonstrated at the Workshop could be rewritten to use X 2567 across the network, so that one could run them from any X-capable device- 2568 -a workstation, an X terminal--and transact with a database across the 2569 network. Although this opens up access a little, assuming one has enough 2570 network to handle it, it does not provide an interface to develop a 2571 program that conveniently integrates information from multiple databases. 2572 X is a viewing technology that has limits. In a real sense, it is just a 2573 graphical version of remote log-in across the network. X-type applications 2574 represent only one step in the progression towards real access. 2575 2576 LYNCH next discussed barriers to the distribution of networked multimedia 2577 information. The heart of the problem is a lack of standards to provide 2578 the ability for computers to talk to each other, retrieve information, 2579 and shuffle it around fairly casually. At the moment, little progress is 2580 being made on standards for networked information; for example, present 2581 standards do not cover images, digital voice, and digital video. A 2582 useful tool kit of exchange formats for basic texts is only now being 2583 assembled. The synchronization of content streams (i.e., synchronizing a 2584 voice track to a video track, establishing temporal relations between 2585 different components in a multimedia object) constitutes another issue 2586 for networked multimedia that is just beginning to receive attention. 2587 2588 Underlying network protocols also need some work; good, real-time 2589 delivery protocols on the Internet do not yet exist. In LYNCH's view, 2590 highly important in this context is the notion of networked digital 2591 object IDs, the ability of one object on the network to point to another 2592 object (or component thereof) on the network. Serious bandwidth issues 2593 also exist. LYNCH was uncertain if billion-bit-per-second networks would 2594 prove sufficient if numerous people ran video in parallel. 2595 2596 LYNCH concluded by offering an issue for database creators to consider, 2597 as well as several comments about what might constitute good trial 2598 multimedia experiments. In a networked information world the database 2599 builder or service builder (publisher) does not exercise the same 2600 extensive control over the integrity of the presentation; strange 2601 programs "munge" with one's data before the user sees it. Serious 2602 thought must be given to what guarantees integrity of presentation. Part 2603 of that is related to where one draws the boundaries around a networked 2604 information service. This question of presentation integrity in 2605 client-server computing has not been stressed enough in the academic 2606 world, LYNCH argued, though commercial service providers deal with it 2607 regularly. 2608 2609 Concerning multimedia, LYNCH observed that good multimedia at the moment 2610 is hideously expensive to produce. He recommended producing multimedia 2611 with either very high sale value, or multimedia with a very long life 2612 span, or multimedia that will have a very broad usage base and whose 2613 costs therefore can be amortized among large numbers of users. In this 2614 connection, historical and humanistically oriented material may be a good 2615 place to start, because it tends to have a longer life span than much of 2616 the scientific material, as well as a wider user base. LYNCH noted, for 2617 example, that American Memory fits many of the criteria outlined. He 2618 remarked the extensive discussion about bringing the Internet or the 2619 National Research and Education Network (NREN) into the K-12 environment 2620 as a way of helping the American educational system. 2621 2622 LYNCH closed by noting that the kinds of applications demonstrated struck 2623 him as excellent justifications of broad-scale networking for K-12, but 2624 that at this time no "killer" application exists to mobilize the K-12 2625 community to obtain connectivity. 2626 2627 ****** 2628 2629 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 2630 DISCUSSION * Dearth of genuinely interesting applications on the network 2631 a slow-changing situation * The issue of the integrity of presentation in 2632 a networked environment * Several reasons why CD-ROM software does not 2633 network * 2634 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 2635 2636 During the discussion period that followed LYNCH's presentation, several 2637 additional points were made. 2638 2639 LYNCH reiterated even more strongly his contention that, historically, 2640 once one goes outside high-end science and the group of those who need 2641 access to supercomputers, there is a great dearth of genuinely 2642 interesting applications on the network. He saw this situation changing 2643 slowly, with some of the scientific databases and scholarly discussion 2644 groups and electronic journals coming on as well as with the availability 2645 of Wide Area Information Servers (WAIS) and some of the databases that 2646 are being mounted there. However, many of those things do not seem to 2647 have piqued great popular interest. For instance, most high school 2648 students of LYNCH's acquaintance would not qualify as devotees of serious 2649 molecular biology. 2650 2651 Concerning the issue of the integrity of presentation, LYNCH believed 2652 that a couple of information providers have laid down the law at least on 2653 certain things. For example, his recollection was that the National 2654 Library of Medicine feels strongly that one needs to employ the 2655 identifier field if he or she is to mount a database commercially. The 2656 problem with a real networked environment is that one does not know who 2657 is reformatting and reprocessing one's data when one enters a client 2658 server mode. It becomes anybody's guess, for example, if the network 2659 uses a Z39.50 server, or what clients are doing with one's data. A data 2660 provider can say that his contract will only permit clients to have 2661 access to his data after he vets them and their presentation and makes 2662 certain it suits him. But LYNCH held out little expectation that the 2663 network marketplace would evolve in that way, because it required too 2664 much prior negotiation. 2665 2666 CD-ROM software does not network for a variety of reasons, LYNCH said. 2667 He speculated that CD-ROM publishers are not eager to have their products 2668 really hook into wide area networks, because they fear it will make their 2669 data suppliers nervous. Moreover, until relatively recently, one had to 2670 be rather adroit to run a full TCP/IP stack plus applications on a 2671 PC-size machine, whereas nowadays it is becoming easier as PCs grow 2672 bigger and faster. LYNCH also speculated that software providers had not 2673 heard from their customers until the last year or so, or had not heard 2674 from enough of their customers. 2675 2676 ****** 2677 2678 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 2679 BESSER * Implications of disseminating images on the network; planning 2680 the distribution of multimedia documents poses two critical 2681 implementation problems * Layered approach represents the way to deal 2682 with users' capabilities * Problems in platform design; file size and its 2683 implications for networking * Transmission of megabyte size images 2684 impractical * Compression and decompression at the user's end * Promising 2685 trends for compression * A disadvantage of using X-Windows * A project at 2686 the Smithsonian that mounts images on several networks * 2687 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 2688 2689 Howard BESSER, School of Library and Information Science, University of 2690 Pittsburgh, spoke primarily about multimedia, focusing on images and the 2691 broad implications of disseminating them on the network. He argued that 2692 planning the distribution of multimedia documents posed two critical 2693 implementation problems, which he framed in the form of two questions: 2694 1) What platform will one use and what hardware and software will users 2695 have for viewing of the material? and 2) How can one deliver a 2696 sufficiently robust set of information in an accessible format in a 2697 reasonable amount of time? Depending on whether network or CD-ROM is the 2698 medium used, this question raises different issues of storage, 2699 compression, and transmission. 2700 2701 Concerning the design of platforms (e.g., sound, gray scale, simple 2702 color, etc.) and the various capabilities users may have, BESSER 2703 maintained that a layered approach was the way to deal with users' 2704 capabilities. A result would be that users with less powerful 2705 workstations would simply have less functionality. He urged members of 2706 the audience to advocate standards and accompanying software that handle 2707 layered functionality across a wide variety of platforms. 2708 2709 BESSER also addressed problems in platform design, namely, deciding how 2710 large a machine to design for situations when the largest number of users 2711 have the lowest level of the machine, and one desires higher 2712 functionality. BESSER then proceeded to the question of file size and 2713 its implications for networking. He discussed still images in the main. 2714 For example, a digital color image that fills the screen of a standard 2715 mega-pel workstation (Sun or Next) will require one megabyte of storage 2716 for an eight-bit image or three megabytes of storage for a true color or 2717 twenty-four-bit image. Lossless compression algorithms (that is, 2718 computational procedures in which no data is lost in the process of 2719 compressing [and decompressing] an image--the exact bit-representation is 2720 maintained) might bring storage down to a third of a megabyte per image, 2721 but not much further than that. The question of size makes it difficult 2722 to fit an appropriately sized set of these images on a single disk or to 2723 transmit them quickly enough on a network. 2724 2725 With these full screen mega-pel images that constitute a third of a 2726 megabyte, one gets 1,000-3,000 full-screen images on a one-gigabyte disk; 2727 a standard CD-ROM represents approximately 60 percent of that. Storing 2728 images the size of a PC screen (just 8 bit color) increases storage 2729 capacity to 4,000-12,000 images per gigabyte; 60 percent of that gives 2730 one the size of a CD-ROM, which in turn creates a major problem. One 2731 cannot have full-screen, full-color images with lossless compression; one 2732 must compress them or use a lower resolution. For megabyte-size images, 2733 anything slower than a T-1 speed is impractical. For example, on a 2734 fifty-six-kilobaud line, it takes three minutes to transfer a 2735 one-megabyte file, if it is not compressed; and this speed assumes ideal 2736 circumstances (no other user contending for network bandwidth). Thus, 2737 questions of disk access, remote display, and current telephone 2738 connection speed make transmission of megabyte-size images impractical. 2739 2740 BESSER then discussed ways to deal with these large images, for example, 2741 compression and decompression at the user's end. In this connection, the 2742 issues of how much one is willing to lose in the compression process and 2743 what image quality one needs in the first place are unknown. But what is 2744 known is that compression entails some loss of data. BESSER urged that 2745 more studies be conducted on image quality in different situations, for 2746 example, what kind of images are needed for what kind of disciplines, and 2747 what kind of image quality is needed for a browsing tool, an intermediate 2748 viewing tool, and archiving. 2749 2750 BESSER remarked two promising trends for compression: from a technical 2751 perspective, algorithms that use what is called subjective redundancy 2752 employ principles from visual psycho-physics to identify and remove 2753 information from the image that the human eye cannot perceive; from an 2754 interchange and interoperability perspective, the JPEG (i.e., Joint 2755 Photographic Experts Group, an ISO standard) compression algorithms also 2756 offer promise. These issues of compression and decompression, BESSER 2757 argued, resembled those raised earlier concerning the design of different 2758 platforms. Gauging the capabilities of potential users constitutes a 2759 primary goal. BESSER advocated layering or separating the images from 2760 the applications that retrieve and display them, to avoid tying them to 2761 particular software. 2762 2763 BESSER detailed several lessons learned from his work at Berkeley with 2764 Imagequery, especially the advantages and disadvantages of using 2765 X-Windows. In the latter category, for example, retrieval is tied 2766 directly to one's data, an intolerable situation in the long run on a 2767 networked system. Finally, BESSER described a project of Jim Wallace at 2768 the Smithsonian Institution, who is mounting images in a extremely 2769 rudimentary way on the Compuserv and Genie networks and is preparing to 2770 mount them on America On Line. Although the average user takes over 2771 thirty minutes to download these images (assuming a fairly fast modem), 2772 nevertheless, images have been downloaded 25,000 times. 2773 2774 BESSER concluded his talk with several comments on the business 2775 arrangement between the Smithsonian and Compuserv. He contended that not 2776 enough is known concerning the value of images. 2777 2778 ****** 2779 2780 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 2781 DISCUSSION * Creating digitized photographic collections nearly 2782 impossible except with large organizations like museums * Need for study 2783 to determine quality of images users will tolerate * 2784 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 2785 2786 During the brief exchange between LESK and BESSER that followed, several 2787 clarifications emerged. 2788 2789 LESK argued that the photographers were far ahead of BESSER: It is 2790 almost impossible to create such digitized photographic collections 2791 except with large organizations like museums, because all the 2792 photographic agencies have been going crazy about this and will not sign 2793 licensing agreements on any sort of reasonable terms. LESK had heard 2794 that National Geographic, for example, had tried to buy the right to use 2795 some image in some kind of educational production for $100 per image, but 2796 the photographers will not touch it. They want accounting and payment 2797 for each use, which cannot be accomplished within the system. BESSER 2798 responded that a consortium of photographers, headed by a former National 2799 Geographic photographer, had started assembling its own collection of 2800 electronic reproductions of images, with the money going back to the 2801 cooperative. 2802 2803 LESK contended that BESSER was unnecessarily pessimistic about multimedia 2804 images, because people are accustomed to low-quality images, particularly 2805 from video. BESSER urged the launching of a study to determine what 2806 users would tolerate, what they would feel comfortable with, and what 2807 absolutely is the highest quality they would ever need. Conceding that 2808 he had adopted a dire tone in order to arouse people about the issue, 2809 BESSER closed on a sanguine note by saying that he would not be in this 2810 business if he did not think that things could be accomplished. 2811 2812 ****** 2813 2814 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 2815 LARSEN * Issues of scalability and modularity * Geometric growth of the 2816 Internet and the role played by layering * Basic functions sustaining 2817 this growth * A library's roles and functions in a network environment * 2818 Effects of implementation of the Z39.50 protocol for information 2819 retrieval on the library system * The trade-off between volumes of data 2820 and its potential usage * A snapshot of current trends * 2821 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 2822 2823 Ronald LARSEN, associate director for information technology, University 2824 of Maryland at College Park, first addressed the issues of scalability 2825 and modularity. He noted the difficulty of anticipating the effects of 2826 orders-of-magnitude growth, reflecting on the twenty years of experience 2827 with the Arpanet and Internet. Recalling the day's demonstrations of 2828 CD-ROM and optical disk material, he went on to ask if the field has yet 2829 learned how to scale new systems to enable delivery and dissemination 2830 across large-scale networks. 2831 2832 LARSEN focused on the geometric growth of the Internet from its inception 2833 circa 1969 to the present, and the adjustments required to respond to 2834 that rapid growth. To illustrate the issue of scalability, LARSEN 2835 considered computer networks as including three generic components: 2836 computers, network communication nodes, and communication media. Each 2837 component scales (e.g., computers range from PCs to supercomputers; 2838 network nodes scale from interface cards in a PC through sophisticated 2839 routers and gateways; and communication media range from 2,400-baud 2840 dial-up facilities through 4.5-Mbps backbone links, and eventually to 2841 multigigabit-per-second communication lines), and architecturally, the 2842 components are organized to scale hierarchically from local area networks 2843 to international-scale networks. Such growth is made possible by 2844 building layers of communication protocols, as BESSER pointed out. 2845 By layering both physically and logically, a sense of scalability is 2846 maintained from local area networks in offices, across campuses, through 2847 bridges, routers, campus backbones, fiber-optic links, etc., up into 2848 regional networks and ultimately into national and international 2849 networks. 2850 2851 LARSEN then illustrated the geometric growth over a two-year period-- 2852 through September 1991--of the number of networks that comprise the 2853 Internet. This growth has been sustained largely by the availability of 2854 three basic functions: electronic mail, file transfer (ftp), and remote 2855 log-on (telnet). LARSEN also reviewed the growth in the kind of traffic 2856 that occurs on the network. Network traffic reflects the joint contributions 2857 of a larger population of users and increasing use per user. Today one sees 2858 serious applications involving moving images across the network--a rarity 2859 ten years ago. LARSEN recalled and concurred with BESSER's main point 2860 that the interesting problems occur at the application level. 2861 2862 LARSEN then illustrated a model of a library's roles and functions in a 2863 network environment. He noted, in particular, the placement of on-line 2864 catalogues onto the network and patrons obtaining access to the library 2865 increasingly through local networks, campus networks, and the Internet. 2866 LARSEN supported LYNCH's earlier suggestion that we need to address 2867 fundamental questions of networked information in order to build 2868 environments that scale in the information sense as well as in the 2869 physical sense. 2870 2871 LARSEN supported the role of the library system as the access point into 2872 the nation's electronic collections. Implementation of the Z39.50 2873 protocol for information retrieval would make such access practical and 2874 feasible. For example, this would enable patrons in Maryland to search 2875 California libraries, or other libraries around the world that are 2876 conformant with Z39.50 in a manner that is familiar to University of 2877 Maryland patrons. This client-server model also supports moving beyond 2878 secondary content into primary content. (The notion of how one links 2879 from secondary content to primary content, LARSEN said, represents a 2880 fundamental problem that requires rigorous thought.) After noting 2881 numerous network experiments in accessing full-text materials, including 2882 projects supporting the ordering of materials across the network, LARSEN 2883 revisited the issue of transmitting high-density, high-resolution color 2884 images across the network and the large amounts of bandwidth they 2885 require. He went on to address the bandwidth and synchronization 2886 problems inherent in sending full-motion video across the network. 2887 2888 LARSEN illustrated the trade-off between volumes of data in bytes or 2889 orders of magnitude and the potential usage of that data. He discussed 2890 transmission rates (particularly, the time it takes to move various forms 2891 of information), and what one could do with a network supporting 2892 multigigabit-per-second transmission. At the moment, the network 2893 environment includes a composite of data-transmission requirements, 2894 volumes and forms, going from steady to bursty (high-volume) and from 2895 very slow to very fast. This aggregate must be considered in the design, 2896 construction, and operation of multigigabyte networks. 2897 2898 LARSEN's objective is to use the networks and library systems now being 2899 constructed to increase access to resources wherever they exist, and 2900 thus, to evolve toward an on-line electronic virtual library. 2901 2902 LARSEN concluded by offering a snapshot of current trends: continuing 2903 geometric growth in network capacity and number of users; slower 2904 development of applications; and glacial development and adoption of 2905 standards. The challenge is to design and develop each new application 2906 system with network access and scalability in mind. 2907 2908 ****** 2909 2910 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 2911 BROWNRIGG * Access to the Internet cannot be taken for granted * Packet 2912 radio and the development of MELVYL in 1980-81 in the Division of Library 2913 Automation at the University of California * Design criteria for packet 2914 radio * A demonstration project in San Diego and future plans * Spread 2915 spectrum * Frequencies at which the radios will run and plans to 2916 reimplement the WAIS server software in the public domain * Need for an 2917 infrastructure of radios that do not move around * 2918 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 2919 2920 Edwin BROWNRIGG, executive director, Memex Research Institute, first 2921 polled the audience in order to seek out regular users of the Internet as 2922 well as those planning to use it some time in the future. With nearly 2923 everybody in the room falling into one category or the other, BROWNRIGG 2924 made a point re access, namely that numerous individuals, especially those 2925 who use the Internet every day, take for granted their access to it, the 2926 speeds with which they are connected, and how well it all works. 2927 However, as BROWNRIGG discovered between 1987 and 1989 in Australia, 2928 if one wants access to the Internet but cannot afford it or has some 2929 physical boundary that prevents her or him from gaining access, it can 2930 be extremely frustrating. He suggested that because of economics and 2931 physical barriers we were beginning to create a world of haves and have-nots 2932 in the process of scholarly communication, even in the United States. 2933 2934 BROWNRIGG detailed the development of MELVYL in academic year 1980-81 in 2935 the Division of Library Automation at the University of California, in 2936 order to underscore the issue of access to the system, which at the 2937 outset was extremely limited. In short, the project needed to build a 2938 network, which at that time entailed use of satellite technology, that is, 2939 putting earth stations on campus and also acquiring some terrestrial links 2940 from the State of California's microwave system. The installation of 2941 satellite links, however, did not solve the problem (which actually 2942 formed part of a larger problem involving politics and financial resources). 2943 For while the project team could get a signal onto a campus, it had no means 2944 of distributing the signal throughout the campus. The solution involved 2945 adopting a recent development in wireless communication called packet radio, 2946 which combined the basic notion of packet-switching with radio. The project 2947 used this technology to get the signal from a point on campus where it 2948 came down, an earth station for example, into the libraries, because it 2949 found that wiring the libraries, especially the older marble buildings, 2950 would cost $2,000-$5,000 per terminal. 2951 2952 BROWNRIGG noted that, ten years ago, the project had neither the public 2953 policy nor the technology that would have allowed it to use packet radio 2954 in any meaningful way. Since then much had changed. He proceeded to 2955 detail research and development of the technology, how it is being 2956 deployed in California, and what direction he thought it would take. 2957 The design criteria are to produce a high-speed, one-time, low-cost, 2958 high-quality, secure, license-free device (packet radio) that one can 2959 plug in and play today, forget about it, and have access to the Internet. 2960 By high speed, BROWNRIGG meant 1 megabyte and 1.5 megabytes. Those units 2961 have been built, he continued, and are in the process of being 2962 type-certified by an independent underwriting laboratory so that they can 2963 be type-licensed by the Federal Communications Commission. As is the 2964 case with citizens band, one will be able to purchase a unit and not have 2965 to worry about applying for a license. 2966 2967 The basic idea, BROWNRIGG elaborated, is to take high-speed radio data 2968 transmission and create a backbone network that at certain strategic 2969 points in the network will "gateway" into a medium-speed packet radio 2970 (i.e., one that runs at 38.4 kilobytes), so that perhaps by 1994-1995 2971 people, like those in the audience for the price of a VCR could purchase 2972 a medium-speed radio for the office or home, have full network connectivity 2973 to the Internet, and partake of all its services, with no need for an FCC 2974 license and no regular bill from the local common carrier. BROWNRIGG 2975 presented several details of a demonstration project currently taking 2976 place in San Diego and described plans, pending funding, to install a 2977 full-bore network in the San Francisco area. This network will have 600 2978 nodes running at backbone speeds, and 100 of these nodes will be libraries, 2979 which in turn will be the gateway ports to the 38.4 kilobyte radios that 2980 will give coverage for the neighborhoods surrounding the libraries. 2981 2982 BROWNRIGG next explained Part 15.247, a new rule within Title 47 of the 2983 Code of Federal Regulations enacted by the FCC in 1985. This rule 2984 challenged the industry, which has only now risen to the occasion, to 2985 build a radio that would run at no more than one watt of output power and 2986 use a fairly exotic method of modulating the radio wave called spread 2987 spectrum. Spread spectrum in fact permits the building of networks so 2988 that numerous data communications can occur simultaneously, without 2989 interfering with each other, within the same wide radio channel. 2990 2991 BROWNRIGG explained that the frequencies at which the radios would run 2992 are very short wave signals. They are well above standard microwave and 2993 radar. With a radio wave that small, one watt becomes a tremendous punch 2994 per bit and thus makes transmission at reasonable speed possible. In 2995 order to minimize the potential for congestion, the project is 2996 undertaking to reimplement software which has been available in the 2997 networking business and is taken for granted now, for example, TCP/IP, 2998 routing algorithms, bridges, and gateways. In addition, the project 2999 plans to take the WAIS server software in the public domain and 3000 reimplement it so that one can have a WAIS server on a Mac instead of a 3001 Unix machine. The Memex Research Institute believes that libraries, in 3002 particular, will want to use the WAIS servers with packet radio. This 3003 project, which has a team of about twelve people, will run through 1993 3004 and will include the 100 libraries already mentioned as well as other 3005 professionals such as those in the medical profession, engineering, and 3006 law. Thus, the need is to create an infrastructure of radios that do not 3007 move around, which, BROWNRIGG hopes, will solve a problem not only for 3008 libraries but for individuals who, by and large today, do not have access 3009 to the Internet from their homes and offices. 3010 3011 ****** 3012 3013 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 3014 DISCUSSION * Project operating frequencies * 3015 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 3016 3017 During a brief discussion period, which also concluded the day's 3018 proceedings, BROWNRIGG stated that the project was operating in four 3019 frequencies. The slow speed is operating at 435 megahertz, and it would 3020 later go up to 920 megahertz. With the high-speed frequency, the 3021 one-megabyte radios will run at 2.4 gigabits, and 1.5 will run at 5.7. 3022 At 5.7, rain can be a factor, but it would have to be tropical rain, 3023 unlike what falls in most parts of the United States. 3024 3025 ****** 3026 3027 SESSION IV. IMAGE CAPTURE, TEXT CAPTURE, OVERVIEW OF TEXT AND 3028 IMAGE STORAGE FORMATS 3029 3030 William HOOTON, vice president of operations, I-NET, moderated this session. 3031 3032 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 3033 KENNEY * Factors influencing development of CXP * Advantages of using 3034 digital technology versus photocopy and microfilm * A primary goal of 3035 CXP; publishing challenges * Characteristics of copies printed * Quality 3036 of samples achieved in image capture * Several factors to be considered 3037 in choosing scanning * Emphasis of CXP on timely and cost-effective 3038 production of black-and-white printed facsimiles * Results of producing 3039 microfilm from digital files * Advantages of creating microfilm * Details 3040 concerning production * Costs * Role of digital technology in library 3041 preservation * 3042 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 3043 3044 Anne KENNEY, associate director, Department of Preservation and 3045 Conservation, Cornell University, opened her talk by observing that the 3046 Cornell Xerox Project (CXP) has been guided by the assumption that the 3047 ability to produce printed facsimiles or to replace paper with paper 3048 would be important, at least for the present generation of users and 3049 equipment. She described three factors that influenced development of 3050 the project: 1) Because the project has emphasized the preservation of 3051 deteriorating brittle books, the quality of what was produced had to be 3052 sufficiently high to return a paper replacement to the shelf. CXP was 3053 only interested in using: 2) a system that was cost-effective, which 3054 meant that it had to be cost-competitive with the processes currently 3055 available, principally photocopy and microfilm, and 3) new or currently 3056 available product hardware and software. 3057 3058 KENNEY described the advantages that using digital technology offers over 3059 both photocopy and microfilm: 1) The potential exists to create a higher 3060 quality reproduction of a deteriorating original than conventional 3061 light-lens technology. 2) Because a digital image is an encoded 3062 representation, it can be reproduced again and again with no resulting 3063 loss of quality, as opposed to the situation with light-lens processes, 3064 in which there is discernible difference between a second and a 3065 subsequent generation of an image. 3) A digital image can be manipulated 3066 in a number of ways to improve image capture; for example, Xerox has 3067 developed a windowing application that enables one to capture a page 3068 containing both text and illustrations in a manner that optimizes the 3069 reproduction of both. (With light-lens technology, one must choose which 3070 to optimize, text or the illustration; in preservation microfilming, the 3071 current practice is to shoot an illustrated page twice, once to highlight 3072 the text and the second time to provide the best capture for the 3073 illustration.) 4) A digital image can also be edited, density levels 3074 adjusted to remove underlining and stains, and to increase legibility for 3075 faint documents. 5) On-screen inspection can take place at the time of 3076 initial setup and adjustments made prior to scanning, factors that 3077 substantially reduce the number of retakes required in quality control. 3078 3079 A primary goal of CXP has been to evaluate the paper output printed on 3080 the Xerox DocuTech, a high-speed printer that produces 600-dpi pages from 3081 scanned images at a rate of 135 pages a minute. KENNEY recounted several 3082 publishing challenges to represent faithful and legible reproductions of 3083 the originals that the 600-dpi copy for the most part successfully 3084 captured. For example, many of the deteriorating volumes in the project 3085 were heavily illustrated with fine line drawings or halftones or came in 3086 languages such as Japanese, in which the buildup of characters comprised 3087 of varying strokes is difficult to reproduce at lower resolutions; a 3088 surprising number of them came with annotations and mathematical 3089 formulas, which it was critical to be able to duplicate exactly. 3090 3091 KENNEY noted that 1) the copies are being printed on paper that meets the 3092 ANSI standards for performance, 2) the DocuTech printer meets the machine 3093 and toner requirements for proper adhesion of print to page, as described 3094 by the National Archives, and thus 3) paper product is considered to be 3095 the archival equivalent of preservation photocopy. 3096 3097 KENNEY then discussed several samples of the quality achieved in the 3098 project that had been distributed in a handout, for example, a copy of a 3099 print-on-demand version of the 1911 Reed lecture on the steam turbine, 3100 which contains halftones, line drawings, and illustrations embedded in 3101 text; the first four loose pages in the volume compared the capture 3102 capabilities of scanning to photocopy for a standard test target, the 3103 IEEE standard 167A 1987 test chart. In all instances scanning proved 3104 superior to photocopy, though only slightly more so in one. 3105 3106 Conceding the simplistic nature of her review of the quality of scanning 3107 to photocopy, KENNEY described it as one representation of the kinds of 3108 settings that could be used with scanning capabilities on the equipment 3109 CXP uses. KENNEY also pointed out that CXP investigated the quality 3110 achieved with binary scanning only, and noted the great promise in gray 3111 scale and color scanning, whose advantages and disadvantages need to be 3112 examined. She argued further that scanning resolutions and file formats 3113 can represent a complex trade-off between the time it takes to capture 3114 material, file size, fidelity to the original, and on-screen display; and 3115 printing and equipment availability. All these factors must be taken 3116 into consideration. 3117 3118 CXP placed primary emphasis on the production in a timely and 3119 cost-effective manner of printed facsimiles that consisted largely of 3120 black-and-white text. With binary scanning, large files may be 3121 compressed efficiently and in a lossless manner (i.e., no data is lost in 3122 the process of compressing [and decompressing] an image--the exact 3123 bit-representation is maintained) using Group 4 CCITT (i.e., the French 3124 acronym for International Consultative Committee for Telegraph and 3125 Telephone) compression. CXP was getting compression ratios of about 3126 forty to one. Gray-scale compression, which primarily uses JPEG, is much 3127 less economical and can represent a lossy compression (i.e., not 3128 lossless), so that as one compresses and decompresses, the illustration 3129 is subtly changed. While binary files produce a high-quality printed 3130 version, it appears 1) that other combinations of spatial resolution with 3131 gray and/or color hold great promise as well, and 2) that gray scale can 3132 represent a tremendous advantage for on-screen viewing. The quality 3133 associated with binary and gray scale also depends on the equipment used. 3134 For instance, binary scanning produces a much better copy on a binary 3135 printer. 3136 3137 Among CXP's findings concerning the production of microfilm from digital 3138 files, KENNEY reported that the digital files for the same Reed lecture 3139 were used to produce sample film using an electron beam recorder. The 3140 resulting film was faithful to the image capture of the digital files, 3141 and while CXP felt that the text and image pages represented in the Reed 3142 lecture were superior to that of the light-lens film, the resolution 3143 readings for the 600 dpi were not as high as standard microfilming. 3144 KENNEY argued that the standards defined for light-lens technology are 3145 not totally transferable to a digital environment. Moreover, they are 3146 based on definition of quality for a preservation copy. Although making 3147 this case will prove to be a long, uphill struggle, CXP plans to continue 3148 to investigate the issue over the course of the next year. 3149 3150 KENNEY concluded this portion of her talk with a discussion of the 3151 advantages of creating film: it can serve as a primary backup and as a 3152 preservation master to the digital file; it could then become the print 3153 or production master and service copies could be paper, film, optical 3154 disks, magnetic media, or on-screen display. 3155 3156 Finally, KENNEY presented details re production: 3157 3158 * Development and testing of a moderately-high resolution production 3159 scanning workstation represented a third goal of CXP; to date, 1,000 3160 volumes have been scanned, or about 300,000 images. 3161 3162 * The resulting digital files are stored and used to produce 3163 hard-copy replacements for the originals and additional prints on 3164 demand; although the initial costs are high, scanning technology 3165 offers an affordable means for reformatting brittle material. 3166 3167 * A technician in production mode can scan 300 pages per hour when 3168 performing single-sheet scanning, which is a necessity when working 3169 with truly brittle paper; this figure is expected to increase 3170 significantly with subsequent iterations of the software from Xerox; 3171 a three-month time-and-cost study of scanning found that the average 3172 300-page book would take about an hour and forty minutes to scan 3173 (this figure included the time for setup, which involves keying in 3174 primary bibliographic data, going into quality control mode to 3175 define page size, establishing front-to-back registration, and 3176 scanning sample pages to identify a default range of settings for 3177 the entire book--functions not dissimilar to those performed by 3178 filmers or those preparing a book for photocopy). 3179 3180 * The final step in the scanning process involved rescans, which 3181 happily were few and far between, representing well under 1 percent 3182 of the total pages scanned. 3183 3184 In addition to technician time, CXP costed out equipment, amortized over 3185 four years, the cost of storing and refreshing the digital files every 3186 four years, and the cost of printing and binding, book-cloth binding, a 3187 paper reproduction. The total amounted to a little under $65 per single 3188 300-page volume, with 30 percent overhead included--a figure competitive 3189 with the prices currently charged by photocopy vendors. 3190 3191 Of course, with scanning, in addition to the paper facsimile, one is left 3192 with a digital file from which subsequent copies of the book can be 3193 produced for a fraction of the cost of photocopy, with readers afforded 3194 choices in the form of these copies. 3195 3196 KENNEY concluded that digital technology offers an electronic means for a 3197 library preservation effort to pay for itself. If a brittle-book program 3198 included the means of disseminating reprints of books that are in demand 3199 by libraries and researchers alike, the initial investment in capture 3200 could be recovered and used to preserve additional but less popular 3201 books. She disclosed that an economic model for a self-sustaining 3202 program could be developed for CXP's report to the Commission on 3203 Preservation and Access (CPA). 3204 3205 KENNEY stressed that the focus of CXP has been on obtaining high quality 3206 in a production environment. The use of digital technology is viewed as 3207 an affordable alternative to other reformatting options. 3208 3209 ****** 3210 3211 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 3212 ANDRE * Overview and history of NATDP * Various agricultural CD-ROM 3213 products created inhouse and by service bureaus * Pilot project on 3214 Internet transmission * Additional products in progress * 3215 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 3216 3217 Pamela ANDRE, associate director for automation, National Agricultural 3218 Text Digitizing Program (NATDP), National Agricultural Library (NAL), 3219 presented an overview of NATDP, which has been underway at NAL the last 3220 four years, before Judith ZIDAR discussed the technical details. ANDRE 3221 defined agricultural information as a broad range of material going from 3222 basic and applied research in the hard sciences to the one-page pamphlets 3223 that are distributed by the cooperative state extension services on such 3224 things as how to grow blueberries. 3225 3226 NATDP began in late 1986 with a meeting of representatives from the 3227 land-grant library community to deal with the issue of electronic 3228 information. NAL and forty-five of these libraries banded together to 3229 establish this project--to evaluate the technology for converting what 3230 were then source documents in paper form into electronic form, to provide 3231 access to that digital information, and then to distribute it. 3232 Distributing that material to the community--the university community as 3233 well as the extension service community, potentially down to the county 3234 level--constituted the group's chief concern. 3235 3236 Since January 1988 (when the microcomputer-based scanning system was 3237 installed at NAL), NATDP has done a variety of things, concerning which 3238 ZIDAR would provide further details. For example, the first technology 3239 considered in the project's discussion phase was digital videodisc, which 3240 indicates how long ago it was conceived. 3241 3242 Over the four years of this project, four separate CD-ROM products on 3243 four different agricultural topics were created, two at a 3244 scanning-and-OCR station installed at NAL, and two by service bureaus. 3245 Thus, NATDP has gained comparative information in terms of those relative 3246 costs. Each of these products contained the full ASCII text as well as 3247 page images of the material, or between 4,000 and 6,000 pages of material 3248 on these disks. Topics included aquaculture, food, agriculture and 3249 science (i.e., international agriculture and research), acid rain, and 3250 Agent Orange, which was the final product distributed (approximately 3251 eighteen months before the Workshop). 3252 3253 The third phase of NATDP focused on delivery mechanisms other than 3254 CD-ROM. At the suggestion of Clifford LYNCH, who was a technical 3255 consultant to the project at this point, NATDP became involved with the 3256 Internet and initiated a project with the help of North Carolina State 3257 University, in which fourteen of the land-grant university libraries are 3258 transmitting digital images over the Internet in response to interlibrary 3259 loan requests--a topic for another meeting. At this point, the pilot 3260 project had been completed for about a year and the final report would be 3261 available shortly after the Workshop. In the meantime, the project's 3262 success had led to its extension. (ANDRE noted that one of the first 3263 things done under the program title was to select a retrieval package to 3264 use with subsequent products; Windows Personal Librarian was the package 3265 of choice after a lengthy evaluation.) 3266 3267 Three additional products had been planned and were in progress: 3268 3269 1) An arrangement with the American Society of Agronomy--a 3270 professional society that has published the Agronomy Journal since 3271 about 1908--to scan and create bit-mapped images of its journal. 3272 ASA granted permission first to put and then to distribute this 3273 material in electronic form, to hold it at NAL, and to use these 3274 electronic images as a mechanism to deliver documents or print out 3275 material for patrons, among other uses. Effectively, NAL has the 3276 right to use this material in support of its program. 3277 (Significantly, this arrangement offers a potential cooperative 3278 model for working with other professional societies in agriculture 3279 to try to do the same thing--put the journals of particular interest 3280 to agriculture research into electronic form.) 3281 3282 2) An extension of the earlier product on aquaculture. 3283 3284 3) The George Washington Carver Papers--a joint project with 3285 Tuskegee University to scan and convert from microfilm some 3,500 3286 images of Carver's papers, letters, and drawings. 3287 3288 It was anticipated that all of these products would appear no more than 3289 six months after the Workshop. 3290 3291 ****** 3292 3293 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 3294 ZIDAR * (A separate arena for scanning) * Steps in creating a database * 3295 Image capture, with and without performing OCR * Keying in tracking data 3296 * Scanning, with electronic and manual tracking * Adjustments during 3297 scanning process * Scanning resolutions * Compression * De-skewing and 3298 filtering * Image capture from microform: the papers and letters of 3299 George Washington Carver * Equipment used for a scanning system * 3300 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 3301 3302 Judith ZIDAR, coordinator, National Agricultural Text Digitizing Program 3303 (NATDP), National Agricultural Library (NAL), illustrated the technical 3304 details of NATDP, including her primary responsibility, scanning and 3305 creating databases on a topic and putting them on CD-ROM. 3306 3307 (ZIDAR remarked a separate arena from the CD-ROM projects, although the 3308 processing of the material is nearly identical, in which NATDP is also 3309 scanning material and loading it on a Next microcomputer, which in turn 3310 is linked to NAL's integrated library system. Thus, searches in NAL's 3311 bibliographic database will enable people to pull up actual page images 3312 and text for any documents that have been entered.) 3313 3314 In accordance with the session's topic, ZIDAR focused her illustrated 3315 talk on image capture, offering a primer on the three main steps in the 3316 process: 1) assemble the printed publications; 2) design the database 3317 (database design occurs in the process of preparing the material for 3318 scanning; this step entails reviewing and organizing the material, 3319 defining the contents--what will constitute a record, what kinds of 3320 fields will be captured in terms of author, title, etc.); 3) perform a 3321 certain amount of markup on the paper publications. NAL performs this 3322 task record by record, preparing work sheets or some other sort of 3323 tracking material and designing descriptors and other enhancements to be 3324 added to the data that will not be captured from the printed publication. 3325 Part of this process also involves determining NATDP's file and directory 3326 structure: NATDP attempts to avoid putting more than approximately 100 3327 images in a directory, because placing more than that on a CD-ROM would 3328 reduce the access speed. 3329 3330 This up-front process takes approximately two weeks for a 3331 6,000-7,000-page database. The next step is to capture the page images. 3332 How long this process takes is determined by the decision whether or not 3333 to perform OCR. Not performing OCR speeds the process, whereas text 3334 capture requires greater care because of the quality of the image: it 3335 has to be straighter and allowance must be made for text on a page, not 3336 just for the capture of photographs. 3337 3338 NATDP keys in tracking data, that is, a standard bibliographic record 3339 including the title of the book and the title of the chapter, which will 3340 later either become the access information or will be attached to the 3341 front of a full-text record so that it is searchable. 3342 3343 Images are scanned from a bound or unbound publication, chiefly from 3344 bound publications in the case of NATDP, however, because often they are 3345 the only copies and the publications are returned to the shelves. NATDP 3346 usually scans one record at a time, because its database tracking system 3347 tracks the document in that way and does not require further logical 3348 separating of the images. After performing optical character 3349 recognition, NATDP moves the images off the hard disk and maintains a 3350 volume sheet. Though the system tracks electronically, all the 3351 processing steps are also tracked manually with a log sheet. 3352 3353 ZIDAR next illustrated the kinds of adjustments that one can make when 3354 scanning from paper and microfilm, for example, redoing images that need 3355 special handling, setting for dithering or gray scale, and adjusting for 3356 brightness or for the whole book at one time. 3357 3358 NATDP is scanning at 300 dots per inch, a standard scanning resolution. 3359 Though adequate for capturing text that is all of a standard size, 300 3360 dpi is unsuitable for any kind of photographic material or for very small 3361 text. Many scanners allow for different image formats, TIFF, of course, 3362 being a de facto standard. But if one intends to exchange images with 3363 other people, the ability to scan other image formats, even if they are 3364 less common, becomes highly desirable. 3365 3366 CCITT Group 4 is the standard compression for normal black-and-white 3367 images, JPEG for gray scale or color. ZIDAR recommended 1) using the 3368 standard compressions, particularly if one attempts to make material 3369 available and to allow users to download images and reuse them from 3370 CD-ROMs; and 2) maintaining the ability to output an uncompressed image, 3371 because in image exchange uncompressed images are more likely to be able 3372 to cross platforms. 3373 3374 ZIDAR emphasized the importance of de-skewing and filtering as 3375 requirements on NATDP's upgraded system. For instance, scanning bound 3376 books, particularly books published by the federal government whose pages 3377 are skewed, and trying to scan them straight if OCR is to be performed, 3378 is extremely time-consuming. The same holds for filtering of 3379 poor-quality or older materials. 3380 3381 ZIDAR described image capture from microform, using as an example three 3382 reels from a sixty-seven-reel set of the papers and letters of George 3383 Washington Carver that had been produced by Tuskegee University. These 3384 resulted in approximately 3,500 images, which NATDP had had scanned by 3385 its service contractor, Science Applications International Corporation 3386 (SAIC). NATDP also created bibliographic records for access. (NATDP did 3387 not have such specialized equipment as a microfilm scanner. 3388 3389 Unfortunately, the process of scanning from microfilm was not an 3390 unqualified success, ZIDAR reported: because microfilm frame sizes vary, 3391 occasionally some frames were missed, which without spending much time 3392 and money could not be recaptured. 3393 3394 OCR could not be performed from the scanned images of the frames. The 3395 bleeding in the text simply output text, when OCR was run, that could not 3396 even be edited. NATDP tested for negative versus positive images, 3397 landscape versus portrait orientation, and single- versus dual-page 3398 microfilm, none of which seemed to affect the quality of the image; but 3399 also on none of them could OCR be performed. 3400 3401 In selecting the microfilm they would use, therefore, NATDP had other 3402 factors in mind. ZIDAR noted two factors that influenced the quality of 3403 the images: 1) the inherent quality of the original and 2) the amount of 3404 size reduction on the pages. 3405 3406 The Carver papers were selected because they are informative and visually 3407 interesting, treat a single subject, and are valuable in their own right. 3408 The images were scanned and divided into logical records by SAIC, then 3409 delivered, and loaded onto NATDP's system, where bibliographic 3410 information taken directly from the images was added. Scanning was 3411 completed in summer 1991 and by the end of summer 1992 the disk was 3412 scheduled to be published. 3413 3414 Problems encountered during processing included the following: Because 3415 the microfilm scanning had to be done in a batch, adjustment for 3416 individual page variations was not possible. The frame size varied on 3417 account of the nature of the material, and therefore some of the frames 3418 were missed while others were just partial frames. The only way to go 3419 back and capture this material was to print out the page with the 3420 microfilm reader from the missing frame and then scan it in from the 3421 page, which was extremely time-consuming. The quality of the images 3422 scanned from the printout of the microfilm compared unfavorably with that 3423 of the original images captured directly from the microfilm. The 3424 inability to perform OCR also was a major disappointment. At the time, 3425 computer output microfilm was unavailable to test. 3426 3427 The equipment used for a scanning system was the last topic addressed by 3428 ZIDAR. The type of equipment that one would purchase for a scanning 3429 system included: a microcomputer, at least a 386, but preferably a 486; 3430 a large hard disk, 380 megabyte at minimum; a multi-tasking operating 3431 system that allows one to run some things in batch in the background 3432 while scanning or doing text editing, for example, Unix or OS/2 and, 3433 theoretically, Windows; a high-speed scanner and scanning software that 3434 allows one to make the various adjustments mentioned earlier; a 3435 high-resolution monitor (150 dpi ); OCR software and hardware to perform 3436 text recognition; an optical disk subsystem on which to archive all the 3437 images as the processing is done; file management and tracking software. 3438 3439 ZIDAR opined that the software one purchases was more important than the 3440 hardware and might also cost more than the hardware, but it was likely to 3441 prove critical to the success or failure of one's system. In addition to 3442 a stand-alone scanning workstation for image capture, then, text capture 3443 requires one or two editing stations networked to this scanning station 3444 to perform editing. Editing the text takes two or three times as long as 3445 capturing the images. 3446 3447 Finally, ZIDAR stressed the importance of buying an open system that allows 3448 for more than one vendor, complies with standards, and can be upgraded. 3449 3450 ****** 3451 3452 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 3453 WATERS *Yale University Library's master plan to convert microfilm to 3454 digital imagery (POB) * The place of electronic tools in the library of 3455 the future * The uses of images and an image library * Primary input from 3456 preservation microfilm * Features distinguishing POB from CXP and key 3457 hypotheses guiding POB * Use of vendor selection process to facilitate 3458 organizational work * Criteria for selecting vendor * Finalists and 3459 results of process for Yale * Key factor distinguishing vendors * 3460 Components, design principles, and some estimated costs of POB * Role of 3461 preservation materials in developing imaging market * Factors affecting 3462 quality and cost * Factors affecting the usability of complex documents 3463 in image form * 3464 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 3465 3466 Donald WATERS, head of the Systems Office, Yale University Library, 3467 reported on the progress of a master plan for a project at Yale to 3468 convert microfilm to digital imagery, Project Open Book (POB). Stating 3469 that POB was in an advanced stage of planning, WATERS detailed, in 3470 particular, the process of selecting a vendor partner and several key 3471 issues under discussion as Yale prepares to move into the project itself. 3472 He commented first on the vision that serves as the context of POB and 3473 then described its purpose and scope. 3474 3475 WATERS sees the library of the future not necessarily as an electronic 3476 library but as a place that generates, preserves, and improves for its 3477 clients ready access to both intellectual and physical recorded 3478 knowledge. Electronic tools must find a place in the library in the 3479 context of this vision. Several roles for electronic tools include 3480 serving as: indirect sources of electronic knowledge or as "finding" 3481 aids (the on-line catalogues, the article-level indices, registers for 3482 documents and archives); direct sources of recorded knowledge; full-text 3483 images; and various kinds of compound sources of recorded knowledge (the 3484 so-called compound documents of Hypertext, mixed text and image, 3485 mixed-text image format, and multimedia). 3486 3487 POB is looking particularly at images and an image library, the uses to 3488 which images will be put (e.g., storage, printing, browsing, and then use 3489 as input for other processes), OCR as a subsequent process to image 3490 capture, or creating an image library, and also possibly generating 3491 microfilm. 3492 3493 While input will come from a variety of sources, POB is considering 3494 especially input from preservation microfilm. A possible outcome is that 3495 the film and paper which provide the input for the image library 3496 eventually may go off into remote storage, and that the image library may 3497 be the primary access tool. 3498 3499 The purpose and scope of POB focus on imaging. Though related to CXP, 3500 POB has two features which distinguish it: 1) scale--conversion of 3501 10,000 volumes into digital image form; and 2) source--conversion from 3502 microfilm. Given these features, several key working hypotheses guide 3503 POB, including: 1) Since POB is using microfilm, it is not concerned with 3504 the image library as a preservation medium. 2) Digital imagery can improve 3505 access to recorded knowledge through printing and network distribution at 3506 a modest incremental cost of microfilm. 3) Capturing and storing documents 3507 in a digital image form is necessary to further improvements in access. 3508 (POB distinguishes between the imaging, digitizing process and OCR, 3509 which at this stage it does not plan to perform.) 3510 3511 Currently in its first or organizational phase, POB found that it could 3512 use a vendor selection process to facilitate a good deal of the 3513 organizational work (e.g., creating a project team and advisory board, 3514 confirming the validity of the plan, establishing the cost of the project 3515 and a budget, selecting the materials to convert, and then raising the 3516 necessary funds). 3517 3518 POB developed numerous selection criteria, including: a firm committed 3519 to image-document management, the ability to serve as systems integrator 3520 in a large-scale project over several years, interest in developing the 3521 requisite software as a standard rather than a custom product, and a 3522 willingness to invest substantial resources in the project itself. 3523 3524 Two vendors, DEC and Xerox, were selected as finalists in October 1991, 3525 and with the support of the Commission on Preservation and Access, each 3526 was commissioned to generate a detailed requirements analysis for the 3527 project and then to submit a formal proposal for the completion of the 3528 project, which included a budget and costs. The terms were that POB would 3529 pay the loser. The results for Yale of involving a vendor included: 3530 broad involvement of Yale staff across the board at a relatively low 3531 cost, which may have long-term significance in carrying out the project 3532 (twenty-five to thirty university people are engaged in POB); better 3533 understanding of the factors that affect corporate response to markets 3534 for imaging products; a competitive proposal; and a more sophisticated 3535 view of the imaging markets. 3536 3537 The most important factor that distinguished the vendors under 3538 consideration was their identification with the customer. The size and 3539 internal complexity of the company also was an important factor. POB was 3540 looking at large companies that had substantial resources. In the end, 3541 the process generated for Yale two competitive proposals, with Xerox's 3542 the clear winner. WATERS then described the components of the proposal, 3543 the design principles, and some of the costs estimated for the process. 3544 3545 Components are essentially four: a conversion subsystem, a 3546 network-accessible storage subsystem for 10,000 books (and POB expects 3547 200 to 600 dpi storage), browsing stations distributed on the campus 3548 network, and network access to the image printers. 3549 3550 Among the design principles, POB wanted conversion at the highest 3551 possible resolution. Assuming TIFF files, TIFF files with Group 4 3552 compression, TCP/IP, and ethernet network on campus, POB wanted a 3553 client-server approach with image documents distributed to the 3554 workstations and made accessible through native workstation interfaces 3555 such as Windows. POB also insisted on a phased approach to 3556 implementation: 1) a stand-alone, single-user, low-cost entry into the 3557 business with a workstation focused on conversion and allowing POB to 3558 explore user access; 2) movement into a higher-volume conversion with 3559 network-accessible storage and multiple access stations; and 3) a 3560 high-volume conversion, full-capacity storage, and multiple browsing 3561 stations distributed throughout the campus. 3562 3563 The costs proposed for start-up assumed the existence of the Yale network 3564 and its two DocuTech image printers. Other start-up costs are estimated 3565 at $1 million over the three phases. At the end of the project, the annual 3566 operating costs estimated primarily for the software and hardware proposed 3567 come to about $60,000, but these exclude costs for labor needed in the 3568 conversion process, network and printer usage, and facilities management. 3569 3570 Finally, the selection process produced for Yale a more sophisticated 3571 view of the imaging markets: the management of complex documents in 3572 image form is not a preservation problem, not a library problem, but a 3573 general problem in a broad, general industry. Preservation materials are 3574 useful for developing that market because of the qualities of the 3575 material. For example, much of it is out of copyright. The resolution 3576 of key issues such as the quality of scanning and image browsing also 3577 will affect development of that market. 3578 3579 The technology is readily available but changing rapidly. In this 3580 context of rapid change, several factors affect quality and cost, to 3581 which POB intends to pay particular attention, for example, the various 3582 levels of resolution that can be achieved. POB believes it can bring 3583 resolution up to 600 dpi, but an interpolation process from 400 to 600 is 3584 more likely. The variation quality in microfilm will prove to be a 3585 highly important factor. POB may reexamine the standards used to film in 3586 the first place by looking at this process as a follow-on to microfilming. 3587 3588 Other important factors include: the techniques available to the 3589 operator for handling material, the ways of integrating quality control 3590 into the digitizing work flow, and a work flow that includes indexing and 3591 storage. POB's requirement was to be able to deal with quality control 3592 at the point of scanning. Thus, thanks to Xerox, POB anticipates having 3593 a mechanism which will allow it not only to scan in batch form, but to 3594 review the material as it goes through the scanner and control quality 3595 from the outset. 3596 3597 The standards for measuring quality and costs depend greatly on the uses 3598 of the material, including subsequent OCR, storage, printing, and 3599 browsing. But especially at issue for POB is the facility for browsing. 3600 This facility, WATERS said, is perhaps the weakest aspect of imaging 3601 technology and the most in need of development. 3602 3603 A variety of factors affect the usability of complex documents in image 3604 form, among them: 1) the ability of the system to handle the full range 3605 of document types, not just monographs but serials, multi-part 3606 monographs, and manuscripts; 2) the location of the database of record 3607 for bibliographic information about the image document, which POB wants 3608 to enter once and in the most useful place, the on-line catalog; 3) a 3609 document identifier for referencing the bibliographic information in one 3610 place and the images in another; 4) the technique for making the basic 3611 internal structure of the document accessible to the reader; and finally, 3612 5) the physical presentation on the CRT of those documents. POB is ready 3613 to complete this phase now. One last decision involves deciding which 3614 material to scan. 3615 3616 ****** 3617 3618 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 3619 DISCUSSION * TIFF files constitute de facto standard * NARA's experience 3620 with image conversion software and text conversion * RFC 1314 * 3621 Considerable flux concerning available hardware and software solutions * 3622 NAL through-put rate during scanning * Window management questions * 3623 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 3624 3625 In the question-and-answer period that followed WATERS's presentation, 3626 the following points emerged: 3627 3628 * ZIDAR's statement about using TIFF files as a standard meant de 3629 facto standard. This is what most people use and typically exchange 3630 with other groups, across platforms, or even occasionally across 3631 display software. 3632 3633 * HOLMES commented on the unsuccessful experience of NARA in 3634 attempting to run image-conversion software or to exchange between 3635 applications: What are supposedly TIFF files go into other software 3636 that is supposed to be able to accept TIFF but cannot recognize the 3637 format and cannot deal with it, and thus renders the exchange 3638 useless. Re text conversion, he noted the different recognition 3639 rates obtained by substituting the make and model of scanners in 3640 NARA's recent test of an "intelligent" character-recognition product 3641 for a new company. In the selection of hardware and software, 3642 HOLMES argued, software no longer constitutes the overriding factor 3643 it did until about a year ago; rather it is perhaps important to 3644 look at both now. 3645 3646 * Danny Cohen and Alan Katz of the University of Southern California 3647 Information Sciences Institute began circulating as an Internet RFC 3648 (RFC 1314) about a month ago a standard for a TIFF interchange 3649 format for Internet distribution of monochrome bit-mapped images, 3650 which LYNCH said he believed would be used as a de facto standard. 3651 3652 * FLEISCHHAUER's impression from hearing these reports and thinking 3653 about AM's experience was that there is considerable flux concerning 3654 available hardware and software solutions. HOOTON agreed and 3655 commented at the same time on ZIDAR's statement that the equipment 3656 employed affects the results produced. One cannot draw a complete 3657 conclusion by saying it is difficult or impossible to perform OCR 3658 from scanning microfilm, for example, with that device, that set of 3659 parameters, and system requirements, because numerous other people 3660 are accomplishing just that, using other components, perhaps. 3661 HOOTON opined that both the hardware and the software were highly 3662 important. Most of the problems discussed today have been solved in 3663 numerous different ways by other people. Though it is good to be 3664 cognizant of various experiences, this is not to say that it will 3665 always be thus. 3666 3667 * At NAL, the through-put rate of the scanning process for paper, 3668 page by page, performing OCR, ranges from 300 to 600 pages per day; 3669 not performing OCR is considerably faster, although how much faster 3670 is not known. This is for scanning from bound books, which is much 3671 slower. 3672 3673 * WATERS commented on window management questions: DEC proposed an 3674 X-Windows solution which was problematical for two reasons. One was 3675 POB's requirement to be able to manipulate images on the workstation 3676 and bring them down to the workstation itself and the other was 3677 network usage. 3678 3679 ****** 3680 3681 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 3682 THOMA * Illustration of deficiencies in scanning and storage process * 3683 Image quality in this process * Different costs entailed by better image 3684 quality * Techniques for overcoming various de-ficiencies: fixed 3685 thresholding, dynamic thresholding, dithering, image merge * Page edge 3686 effects * 3687 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 3688 3689 George THOMA, chief, Communications Engineering Branch, National Library 3690 of Medicine (NLM), illustrated several of the deficiencies discussed by 3691 the previous speakers. He introduced the topic of special problems by 3692 noting the advantages of electronic imaging. For example, it is regenerable 3693 because it is a coded file, and real-time quality control is possible with 3694 electronic capture, whereas in photographic capture it is not. 3695 3696 One of the difficulties discussed in the scanning and storage process was 3697 image quality which, without belaboring the obvious, means different 3698 things for maps, medical X-rays, or broadcast television. In the case of 3699 documents, THOMA said, image quality boils down to legibility of the 3700 textual parts, and fidelity in the case of gray or color photo print-type 3701 material. Legibility boils down to scan density, the standard in most 3702 cases being 300 dpi. Increasing the resolution with scanners that 3703 perform 600 or 1200 dpi, however, comes at a cost. 3704 3705 Better image quality entails at least four different kinds of costs: 1) 3706 equipment costs, because the CCD (i.e., charge-couple device) with 3707 greater number of elements costs more; 2) time costs that translate to 3708 the actual capture costs, because manual labor is involved (the time is 3709 also dependent on the fact that more data has to be moved around in the 3710 machine in the scanning or network devices that perform the scanning as 3711 well as the storage); 3) media costs, because at high resolutions larger 3712 files have to be stored; and 4) transmission costs, because there is just 3713 more data to be transmitted. 3714 3715 But while resolution takes care of the issue of legibility in image 3716 quality, other deficiencies have to do with contrast and elements on the 3717 page scanned or the image that needed to be removed or clarified. Thus, 3718 THOMA proceeded to illustrate various deficiencies, how they are 3719 manifested, and several techniques to overcome them. 3720 3721 Fixed thresholding was the first technique described, suitable for 3722 black-and-white text, when the contrast does not vary over the page. One 3723 can have many different threshold levels in scanning devices. Thus, 3724 THOMA offered an example of extremely poor contrast, which resulted from 3725 the fact that the stock was a heavy red. This is the sort of image that 3726 when microfilmed fails to provide any legibility whatsoever. Fixed 3727 thresholding is the way to change the black-to-red contrast to the 3728 desired black-to-white contrast. 3729 3730 Other examples included material that had been browned or yellowed by 3731 age. This was also a case of contrast deficiency, and correction was 3732 done by fixed thresholding. A final example boils down to the same 3733 thing, slight variability, but it is not significant. Fixed thresholding 3734 solves this problem as well. The microfilm equivalent is certainly legible, 3735 but it comes with dark areas. Though THOMA did not have a slide of the 3736 microfilm in this case, he did show the reproduced electronic image. 3737 3738 When one has variable contrast over a page or the lighting over the page 3739 area varies, especially in the case where a bound volume has light 3740 shining on it, the image must be processed by a dynamic thresholding 3741 scheme. One scheme, dynamic averaging, allows the threshold level not to 3742 be fixed but to be recomputed for every pixel from the neighboring 3743 characteristics. The neighbors of a pixel determine where the threshold 3744 should be set for that pixel. 3745 3746 THOMA showed an example of a page that had been made deficient by a 3747 variety of techniques, including a burn mark, coffee stains, and a yellow 3748 marker. Application of a fixed-thresholding scheme, THOMA argued, might 3749 take care of several deficiencies on the page but not all of them. 3750 Performing the calculation for a dynamic threshold setting, however, 3751 removes most of the deficiencies so that at least the text is legible. 3752 3753 Another problem is representing a gray level with black-and-white pixels 3754 by a process known as dithering or electronic screening. But dithering 3755 does not provide good image quality for pure black-and-white textual 3756 material. THOMA illustrated this point with examples. Although its 3757 suitability for photoprint is the reason for electronic screening or 3758 dithering, it cannot be used for every compound image. In the document 3759 that was distributed by CXP, THOMA noticed that the dithered image of the 3760 IEEE test chart evinced some deterioration in the text. He presented an 3761 extreme example of deterioration in the text in which compounded 3762 documents had to be set right by other techniques. The technique 3763 illustrated by the present example was an image merge in which the page 3764 is scanned twice and the settings go from fixed threshold to the 3765 dithering matrix; the resulting images are merged to give the best 3766 results with each technique. 3767 3768 THOMA illustrated how dithering is also used in nonphotographic or 3769 nonprint materials with an example of a grayish page from a medical text, 3770 which was reproduced to show all of the gray that appeared in the 3771 original. Dithering provided a reproduction of all the gray in the 3772 original of another example from the same text. 3773 3774 THOMA finally illustrated the problem of bordering, or page-edge, 3775 effects. Books and bound volumes that are placed on a photocopy machine 3776 or a scanner produce page-edge effects that are undesirable for two 3777 reasons: 1) the aesthetics of the image; after all, if the image is to 3778 be preserved, one does not necessarily want to keep all of its 3779 deficiencies; 2) compression (with the bordering problem THOMA 3780 illustrated, the compression ratio deteriorated tremendously). One way 3781 to eliminate this more serious problem is to have the operator at the 3782 point of scanning window the part of the image that is desirable and 3783 automatically turn all of the pixels out of that picture to white. 3784 3785 ****** 3786 3787 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 3788 FLEISCHHAUER * AM's experience with scanning bound materials * Dithering 3789 * 3790 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 3791 3792 Carl FLEISCHHAUER, coordinator, American Memory, Library of Congress, 3793 reported AM's experience with scanning bound materials, which he likened 3794 to the problems involved in using photocopying machines. Very few 3795 devices in the industry offer book-edge scanning, let alone book cradles. 3796 The problem may be unsolvable, FLEISCHHAUER said, because a large enough 3797 market does not exist for a preservation-quality scanner. AM is using a 3798 Kurzweil scanner, which is a book-edge scanner now sold by Xerox. 3799 3800 Devoting the remainder of his brief presentation to dithering, 3801 FLEISCHHAUER related AM's experience with a contractor who was using 3802 unsophisticated equipment and software to reduce moire patterns from 3803 printed halftones. AM took the same image and used the dithering 3804 algorithm that forms part of the same Kurzweil Xerox scanner; it 3805 disguised moire patterns much more effectively. 3806 3807 FLEISCHHAUER also observed that dithering produces a binary file which is 3808 useful for numerous purposes, for example, printing it on a laser printer 3809 without having to "re-halftone" it. But it tends to defeat efficient 3810 compression, because the very thing that dithers to reduce moire patterns 3811 also tends to work against compression schemes. AM thought the 3812 difference in image quality was worth it. 3813 3814 ****** 3815 3816 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 3817 DISCUSSION * Relative use as a criterion for POB's selection of books to 3818 be converted into digital form * 3819 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 3820 3821 During the discussion period, WATERS noted that one of the criteria for 3822 selecting books among the 10,000 to be converted into digital image form 3823 would be how much relative use they would receive--a subject still 3824 requiring evaluation. The challenge will be to understand whether 3825 coherent bodies of material will increase usage or whether POB should 3826 seek material that is being used, scan that, and make it more accessible. 3827 POB might decide to digitize materials that are already heavily used, in 3828 order to make them more accessible and decrease wear on them. Another 3829 approach would be to provide a large body of intellectually coherent 3830 material that may be used more in digital form than it is currently used 3831 in microfilm. POB would seek material that was out of copyright. 3832 3833 ****** 3834 3835 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 3836 BARONAS * Origin and scope of AIIM * Types of documents produced in 3837 AIIM's standards program * Domain of AIIM's standardization work * AIIM's 3838 structure * TC 171 and MS23 * Electronic image management standards * 3839 Categories of EIM standardization where AIIM standards are being 3840 developed * 3841 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 3842 3843 Jean BARONAS, senior manager, Department of Standards and Technology, 3844 Association for Information and Image Management (AIIM), described the 3845 not-for-profit association and the national and international programs 3846 for standardization in which AIIM is active. 3847 3848 Accredited for twenty-five years as the nation's standards development 3849 organization for document image management, AIIM began life in a library 3850 community developing microfilm standards. Today the association 3851 maintains both its library and business-image management standardization 3852 activities--and has moved into electronic image-management 3853 standardization (EIM). 3854 3855 BARONAS defined the program's scope. AIIM deals with: 1) the 3856 terminology of standards and of the technology it uses; 2) methods of 3857 measurement for the systems, as well as quality; 3) methodologies for 3858 users to evaluate and measure quality; 4) the features of apparatus used 3859 to manage and edit images; and 5) the procedures used to manage images. 3860 3861 BARONAS noted that three types of documents are produced in the AIIM 3862 standards program: the first two, accredited by the American National 3863 Standards Institute (ANSI), are standards and standard recommended 3864 practices. Recommended practices differ from standards in that they 3865 contain more tutorial information. A technical report is not an ANSI 3866 standard. Because AIIM's policies and procedures for developing 3867 standards are approved by ANSI, its standards are labeled ANSI/AIIM, 3868 followed by the number and title of the standard. 3869 3870 BARONAS then illustrated the domain of AIIM's standardization work. For 3871 example, AIIM is the administrator of the U.S. Technical Advisory Group 3872 (TAG) to the International Standards Organization's (ISO) technical 3873 committee, TC l7l Micrographics and Optical Memories for Document and 3874 Image Recording, Storage, and Use. AIIM officially works through ANSI in 3875 the international standardization process. 3876 3877 BARONAS described AIIM's structure, including its board of directors, its 3878 standards board of twelve individuals active in the image-management 3879 industry, its strategic planning and legal admissibility task forces, and 3880 its National Standards Council, which is comprised of the members of a 3881 number of organizations who vote on every AIIM standard before it is 3882 published. BARONAS pointed out that AIIM's liaisons deal with numerous 3883 other standards developers, including the optical disk community, office 3884 and publishing systems, image-codes-and-character set committees, and the 3885 National Information Standards Organization (NISO). 3886 3887 BARONAS illustrated the procedures of TC l7l, which covers all aspects of 3888 image management. When AIIM's national program has conceptualized a new 3889 project, it is usually submitted to the international level, so that the 3890 member countries of TC l7l can simultaneously work on the development of 3891 the standard or the technical report. BARONAS also illustrated a classic 3892 microfilm standard, MS23, which deals with numerous imaging concepts that 3893 apply to electronic imaging. Originally developed in the l970s, revised 3894 in the l980s, and revised again in l991, this standard is scheduled for 3895 another revision. MS23 is an active standard whereby users may propose 3896 new density ranges and new methods of evaluating film images in the 3897 standard's revision. 3898 3899 BARONAS detailed several electronic image-management standards, for 3900 instance, ANSI/AIIM MS44, a quality-control guideline for scanning 8.5" 3901 by 11" black-and-white office documents. This standard is used with the 3902 IEEE fax image--a continuous tone photographic image with gray scales, 3903 text, and several continuous tone pictures--and AIIM test target number 3904 2, a representative document used in office document management. 3905 3906 BARONAS next outlined the four categories of EIM standardization in which 3907 AIIM standards are being developed: transfer and retrieval, evaluation, 3908 optical disc and document scanning applications, and design and 3909 conversion of documents. She detailed several of the main projects of 3910 each: 1) in the category of image transfer and retrieval, a bi-level 3911 image transfer format, ANSI/AIIM MS53, which is a proposed standard that 3912 describes a file header for image transfer between unlike systems when 3913 the images are compressed using G3 and G4 compression; 2) the category of 3914 image evaluation, which includes the AIIM-proposed TR26 tutorial on image 3915 resolution (this technical report will treat the differences and 3916 similarities between classical or photographic and electronic imaging); 3917 3) design and conversion, which includes a proposed technical report 3918 called "Forms Design Optimization for EIM" (this report considers how 3919 general-purpose business forms can be best designed so that scanning is 3920 optimized; reprographic characteristics such as type, rules, background, 3921 tint, and color will likewise be treated in the technical report); 4) 3922 disk and document scanning applications includes a project a) on planning 3923 platters and disk management, b) on generating an application profile for 3924 EIM when images are stored and distributed on CD-ROM, and c) on 3925 evaluating SCSI2, and how a common command set can be generated for SCSI2 3926 so that document scanners are more easily integrated. (ANSI/AIIM MS53 3927 will also apply to compressed images.) 3928 3929 ****** 3930 3931 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 3932 BATTIN * The implications of standards for preservation * A major 3933 obstacle to successful cooperation * A hindrance to access in the digital 3934 environment * Standards a double-edged sword for those concerned with the 3935 preservation of the human record * Near-term prognosis for reliable 3936 archival standards * Preservation concerns for electronic media * Need 3937 for reconceptualizing our preservation principles * Standards in the real 3938 world and the politics of reproduction * Need to redefine the concept of 3939 archival and to begin to think in terms of life cycles * Cooperation and 3940 the La Guardia Eight * Concerns generated by discussions on the problems 3941 of preserving text and image * General principles to be adopted in a 3942 world without standards * 3943 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 3944 3945 Patricia BATTIN, president, the Commission on Preservation and Access 3946 (CPA), addressed the implications of standards for preservation. She 3947 listed several areas where the library profession and the analog world of 3948 the printed book had made enormous contributions over the past hundred 3949 years--for example, in bibliographic formats, binding standards, and, most 3950 important, in determining what constitutes longevity or archival quality. 3951 3952 Although standards have lightened the preservation burden through the 3953 development of national and international collaborative programs, 3954 nevertheless, a pervasive mistrust of other people's standards remains a 3955 major obstacle to successful cooperation, BATTIN said. 3956 3957 The zeal to achieve perfection, regardless of the cost, has hindered 3958 rather than facilitated access in some instances, and in the digital 3959 environment, where no real standards exist, has brought an ironically 3960 just reward. 3961 3962 BATTIN argued that standards are a double-edged sword for those concerned 3963 with the preservation of the human record, that is, the provision of 3964 access to recorded knowledge in a multitude of media as far into the 3965 future as possible. Standards are essential to facilitate 3966 interconnectivity and access, but, BATTIN said, as LYNCH pointed out 3967 yesterday, if set too soon they can hinder creativity, expansion of 3968 capability, and the broadening of access. The characteristics of 3969 standards for digital imagery differ radically from those for analog 3970 imagery. And the nature of digital technology implies continuing 3971 volatility and change. To reiterate, precipitous standard-setting can 3972 inhibit creativity, but delayed standard-setting results in chaos. 3973 3974 Since in BATTIN'S opinion the near-term prognosis for reliable archival 3975 standards, as defined by librarians in the analog world, is poor, two 3976 alternatives remain: standing pat with the old technology, or 3977 reconceptualizing. 3978 3979 Preservation concerns for electronic media fall into two general domains. 3980 One is the continuing assurance of access to knowledge originally 3981 generated, stored, disseminated, and used in electronic form. This 3982 domain contains several subdivisions, including 1) the closed, 3983 proprietary systems discussed the previous day, bundled information such 3984 as electronic journals and government agency records, and electronically 3985 produced or captured raw data; and 2) the application of digital 3986 technologies to the reformatting of materials originally published on a 3987 deteriorating analog medium such as acid paper or videotape. 3988 3989 The preservation of electronic media requires a reconceptualizing of our 3990 preservation principles during a volatile, standardless transition which 3991 may last far longer than any of us envision today. BATTIN urged the 3992 necessity of shifting focus from assessing, measuring, and setting 3993 standards for the permanence of the medium to the concept of managing 3994 continuing access to information stored on a variety of media and 3995 requiring a variety of ever-changing hardware and software for access--a 3996 fundamental shift for the library profession. 3997 3998 BATTIN offered a primer on how to move forward with reasonable confidence 3999 in a world without standards. Her comments fell roughly into two sections: 4000 1) standards in the real world and 2) the politics of reproduction. 4001 4002 In regard to real-world standards, BATTIN argued the need to redefine the 4003 concept of archive and to begin to think in terms of life cycles. In 4004 the past, the naive assumption that paper would last forever produced a 4005 cavalier attitude toward life cycles. The transient nature of the 4006 electronic media has compelled people to recognize and accept upfront the 4007 concept of life cycles in place of permanency. 4008 4009 Digital standards have to be developed and set in a cooperative context 4010 to ensure efficient exchange of information. Moreover, during this 4011 transition period, greater flexibility concerning how concepts such as 4012 backup copies and archival copies in the CXP are defined is necessary, 4013 or the opportunity to move forward will be lost. 4014 4015 In terms of cooperation, particularly in the university setting, BATTIN 4016 also argued the need to avoid going off in a hundred different 4017 directions. The CPA has catalyzed a small group of universities called 4018 the La Guardia Eight--because La Guardia Airport is where meetings take 4019 place--Harvard, Yale, Cornell, Princeton, Penn State, Tennessee, 4020 Stanford, and USC, to develop a digital preservation consortium to look 4021 at all these issues and develop de facto standards as we move along, 4022 instead of waiting for something that is officially blessed. Continuing 4023 to apply analog values and definitions of standards to the digital 4024 environment, BATTIN said, will effectively lead to forfeiture of the 4025 benefits of digital technology to research and scholarship. 4026 4027 Under the second rubric, the politics of reproduction, BATTIN reiterated 4028 an oft-made argument concerning the electronic library, namely, that it 4029 is more difficult to transform than to create, and nowhere is that belief 4030 expressed more dramatically than in the conversion of brittle books to 4031 new media. Preserving information published in electronic media involves 4032 making sure the information remains accessible and that digital 4033 information is not lost through reproduction. In the analog world of 4034 photocopies and microfilm, the issue of fidelity to the original becomes 4035 paramount, as do issues of "Whose fidelity?" and "Whose original?" 4036 4037 BATTIN elaborated these arguments with a few examples from a recent study 4038 conducted by the CPA on the problems of preserving text and image. 4039 Discussions with scholars, librarians, and curators in a variety of 4040 disciplines dependent on text and image generated a variety of concerns, 4041 for example: 1) Copy what is, not what the technology is capable of. 4042 This is very important for the history of ideas. Scholars wish to know 4043 what the author saw and worked from. And make available at the 4044 workstation the opportunity to erase all the defects and enhance the 4045 presentation. 2) The fidelity of reproduction--what is good enough, what 4046 can we afford, and the difference it makes--issues of subjective versus 4047 objective resolution. 3) The differences between primary and secondary 4048 users. Restricting the definition of primary user to the one in whose 4049 discipline the material has been published runs one headlong into the 4050 reality that these printed books have had a host of other users from a 4051 host of other disciplines, who not only were looking for very different 4052 things, but who also shared values very different from those of the 4053 primary user. 4) The relationship of the standard of reproduction to new 4054 capabilities of scholarship--the browsing standard versus an archival 4055 standard. How good must the archival standard be? Can a distinction be 4056 drawn between potential users in setting standards for reproduction? 4057 Archival storage, use copies, browsing copies--ought an attempt to set 4058 standards even be made? 5) Finally, costs. How much are we prepared to 4059 pay to capture absolute fidelity? What are the trade-offs between vastly 4060 enhanced access, degrees of fidelity, and costs? 4061 4062 These standards, BATTIN concluded, serve to complicate further the 4063 reproduction process, and add to the long list of technical standards 4064 that are necessary to ensure widespread access. Ways to articulate and 4065 analyze the costs that are attached to the different levels of standards 4066 must be found. 4067 4068 Given the chaos concerning standards, which promises to linger for the 4069 foreseeable future, BATTIN urged adoption of the following general 4070 principles: 4071 4072 * Strive to understand the changing information requirements of 4073 scholarly disciplines as more and more technology is integrated into 4074 the process of research and scholarly communication in order to meet 4075 future scholarly needs, not to build for the past. Capture 4076 deteriorating information at the highest affordable resolution, even 4077 though the dissemination and display technologies will lag. 4078 4079 * Develop cooperative mechanisms to foster agreement on protocols 4080 for document structure and other interchange mechanisms necessary 4081 for widespread dissemination and use before official standards are 4082 set. 4083 4084 * Accept that, in a transition period, de facto standards will have 4085 to be developed. 4086 4087 * Capture information in a way that keeps all options open and 4088 provides for total convertibility: OCR, scanning of microfilm, 4089 producing microfilm from scanned documents, etc. 4090 4091 * Work closely with the generators of information and the builders 4092 of networks and databases to ensure that continuing accessibility is 4093 a primary concern from the beginning. 4094 4095 * Piggyback on standards under development for the broad market, and 4096 avoid library-specific standards; work with the vendors, in order to 4097 take advantage of that which is being standardized for the rest of 4098 the world. 4099 4100 * Concentrate efforts on managing permanence in the digital world, 4101 rather than perfecting the longevity of a particular medium. 4102 4103 ****** 4104 4105 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 4106 DISCUSSION * Additional comments on TIFF * 4107 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 4108 4109 During the brief discussion period that followed BATTIN's presentation, 4110 BARONAS explained that TIFF was not developed in collaboration with or 4111 under the auspices of AIIM. TIFF is a company product, not a standard, 4112 is owned by two corporations, and is always changing. BARONAS also 4113 observed that ANSI/AIIM MS53, a bi-level image file transfer format that 4114 allows unlike systems to exchange images, is compatible with TIFF as well 4115 as with DEC's architecture and IBM's MODCA/IOCA. 4116 4117 ****** 4118 4119 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 4120 HOOTON * Several questions to be considered in discussing text conversion 4121 * 4122 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 4123 4124 HOOTON introduced the final topic, text conversion, by noting that it is 4125 becoming an increasingly important part of the imaging business. Many 4126 people now realize that it enhances their system to be able to have more 4127 and more character data as part of their imaging system. Re the issue of 4128 OCR versus rekeying, HOOTON posed several questions: How does one get 4129 text into computer-readable form? Does one use automated processes? 4130 Does one attempt to eliminate the use of operators where possible? 4131 Standards for accuracy, he said, are extremely important: it makes a 4132 major difference in cost and time whether one sets as a standard 98.5 4133 percent acceptance or 99.5 percent. He mentioned outsourcing as a 4134 possibility for converting text. Finally, what one does with the image 4135 to prepare it for the recognition process is also important, he said, 4136 because such preparation changes how recognition is viewed, as well as 4137 facilitates recognition itself. 4138 4139 ****** 4140 4141 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 4142 LESK * Roles of participants in CORE * Data flow * The scanning process * 4143 The image interface * Results of experiments involving the use of 4144 electronic resources and traditional paper copies * Testing the issue of 4145 serendipity * Conclusions * 4146 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 4147 4148 Michael LESK, executive director, Computer Science Research, Bell 4149 Communications Research, Inc. (Bellcore), discussed the Chemical Online 4150 Retrieval Experiment (CORE), a cooperative project involving Cornell 4151 University, OCLC, Bellcore, and the American Chemical Society (ACS). 4152 4153 LESK spoke on 1) how the scanning was performed, including the unusual 4154 feature of page segmentation, and 2) the use made of the text and the 4155 image in experiments. 4156 4157 Working with the chemistry journals (because ACS has been saving its 4158 typesetting tapes since the mid-1970s and thus has a significant back-run 4159 of the most important chemistry journals in the United States), CORE is 4160 attempting to create an automated chemical library. Approximately a 4161 quarter of the pages by square inch are made up of images of 4162 quasi-pictorial material; dealing with the graphic components of the 4163 pages is extremely important. LESK described the roles of participants 4164 in CORE: 1) ACS provides copyright permission, journals on paper, 4165 journals on microfilm, and some of the definitions of the files; 2) at 4166 Bellcore, LESK chiefly performs the data preparation, while Dennis Egan 4167 performs experiments on the users of chemical abstracts, and supplies the 4168 indexing and numerous magnetic tapes; 3) Cornell provides the site of the 4169 experiment; 4) OCLC develops retrieval software and other user interfaces. 4170 Various manufacturers and publishers have furnished other help. 4171 4172 Concerning data flow, Bellcore receives microfilm and paper from ACS; the 4173 microfilm is scanned by outside vendors, while the paper is scanned 4174 inhouse on an Improvision scanner, twenty pages per minute at 300 dpi, 4175 which provides sufficient quality for all practical uses. LESK would 4176 prefer to have more gray level, because one of the ACS journals prints on 4177 some colored pages, which creates a problem. 4178 4179 Bellcore performs all this scanning, creates a page-image file, and also 4180 selects from the pages the graphics, to mix with the text file (which is 4181 discussed later in the Workshop). The user is always searching the ASCII 4182 file, but she or he may see a display based on the ASCII or a display 4183 based on the images. 4184 4185 LESK illustrated how the program performs page analysis, and the image 4186 interface. (The user types several words, is presented with a list-- 4187 usually of the titles of articles contained in an issue--that derives 4188 from the ASCII, clicks on an icon and receives an image that mirrors an 4189 ACS page.) LESK also illustrated an alternative interface, based on text 4190 on the ASCII, the so-called SuperBook interface from Bellcore. 4191 4192 LESK next presented the results of an experiment conducted by Dennis Egan 4193 and involving thirty-six students at Cornell, one third of them 4194 undergraduate chemistry majors, one third senior undergraduate chemistry 4195 majors, and one third graduate chemistry students. A third of them 4196 received the paper journals, the traditional paper copies and chemical 4197 abstracts on paper. A third received image displays of the pictures of 4198 the pages, and a third received the text display with pop-up graphics. 4199 4200 The students were given several questions made up by some chemistry 4201 professors. The questions fell into five classes, ranging from very easy 4202 to very difficult, and included questions designed to simulate browsing 4203 as well as a traditional information retrieval-type task. 4204 4205 LESK furnished the following results. In the straightforward question 4206 search--the question being, what is the phosphorus oxygen bond distance 4207 and hydroxy phosphate?--the students were told that they could take 4208 fifteen minutes and, then, if they wished, give up. The students with 4209 paper took more than fifteen minutes on average, and yet most of them 4210 gave up. The students with either electronic format, text or image, 4211 received good scores in reasonable time, hardly ever had to give up, and 4212 usually found the right answer. 4213 4214 In the browsing study, the students were given a list of eight topics, 4215 told to imagine that an issue of the Journal of the American Chemical 4216 Society had just appeared on their desks, and were also told to flip 4217 through it and to find topics mentioned in the issue. The average scores 4218 were about the same. (The students were told to answer yes or no about 4219 whether or not particular topics appeared.) The errors, however, were 4220 quite different. The students with paper rarely said that something 4221 appeared when it had not. But they often failed to find something 4222 actually mentioned in the issue. The computer people found numerous 4223 things, but they also frequently said that a topic was mentioned when it 4224 was not. (The reason, of course, was that they were performing word 4225 searches. They were finding that words were mentioned and they were 4226 concluding that they had accomplished their task.) 4227 4228 This question also contained a trick to test the issue of serendipity. 4229 The students were given another list of eight topics and instructed, 4230 without taking a second look at the journal, to recall how many of this 4231 new list of eight topics were in this particular issue. This was an 4232 attempt to see if they performed better at remembering what they were not 4233 looking for. They all performed about the same, paper or electronics, 4234 about 62 percent accurate. In short, LESK said, people were not very 4235 good when it came to serendipity, but they were no worse at it with 4236 computers than they were with paper. 4237 4238 (LESK gave a parenthetical illustration of the learning curve of students 4239 who used SuperBook.) 4240 4241 The students using the electronic systems started off worse than the ones 4242 using print, but by the third of the three sessions in the series had 4243 caught up to print. As one might expect, electronics provide a much 4244 better means of finding what one wants to read; reading speeds, once the 4245 object of the search has been found, are about the same. 4246 4247 Almost none of the students could perform the hard task--the analogous 4248 transformation. (It would require the expertise of organic chemists to 4249 complete.) But an interesting result was that the students using the text 4250 search performed terribly, while those using the image system did best. 4251 That the text search system is driven by text offers the explanation. 4252 Everything is focused on the text; to see the pictures, one must press 4253 on an icon. Many students found the right article containing the answer 4254 to the question, but they did not click on the icon to bring up the right 4255 figure and see it. They did not know that they had found the right place, 4256 and thus got it wrong. 4257 4258 The short answer demonstrated by this experiment was that in the event 4259 one does not know what to read, one needs the electronic systems; the 4260 electronic systems hold no advantage at the moment if one knows what to 4261 read, but neither do they impose a penalty. 4262 4263 LESK concluded by commenting that, on one hand, the image system was easy 4264 to use. On the other hand, the text display system, which represented 4265 twenty man-years of work in programming and polishing, was not winning, 4266 because the text was not being read, just searched. The much easier 4267 system is highly competitive as well as remarkably effective for the 4268 actual chemists. 4269 4270 ****** 4271 4272 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 4273 ERWAY * Most challenging aspect of working on AM * Assumptions guiding 4274 AM's approach * Testing different types of service bureaus * AM's 4275 requirement for 99.95 percent accuracy * Requirements for text-coding * 4276 Additional factors influencing AM's approach to coding * Results of AM's 4277 experience with rekeying * Other problems in dealing with service bureaus 4278 * Quality control the most time-consuming aspect of contracting out 4279 conversion * Long-term outlook uncertain * 4280 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 4281 4282 To Ricky ERWAY, associate coordinator, American Memory, Library of 4283 Congress, the constant variety of conversion projects taking place 4284 simultaneously represented perhaps the most challenging aspect of working 4285 on AM. Thus, the challenge was not to find a solution for text 4286 conversion but a tool kit of solutions to apply to LC's varied 4287 collections that need to be converted. ERWAY limited her remarks to the 4288 process of converting text to machine-readable form, and the variety of 4289 LC's text collections, for example, bound volumes, microfilm, and 4290 handwritten manuscripts. 4291 4292 Two assumptions have guided AM's approach, ERWAY said: 1) A desire not 4293 to perform the conversion inhouse. Because of the variety of formats and 4294 types of texts, to capitalize the equipment and have the talents and 4295 skills to operate them at LC would be extremely expensive. Further, the 4296 natural inclination to upgrade to newer and better equipment each year 4297 made it reasonable for AM to focus on what it did best and seek external 4298 conversion services. Using service bureaus also allowed AM to have 4299 several types of operations take place at the same time. 2) AM was not a 4300 technology project, but an effort to improve access to library 4301 collections. Hence, whether text was converted using OCR or rekeying 4302 mattered little to AM. What mattered were cost and accuracy of results. 4303 4304 AM considered different types of service bureaus and selected three to 4305 perform several small tests in order to acquire a sense of the field. 4306 The sample collections with which they worked included handwritten 4307 correspondence, typewritten manuscripts from the 1940s, and 4308 eighteenth-century printed broadsides on microfilm. On none of these 4309 samples was OCR performed; they were all rekeyed. AM had several special 4310 requirements for the three service bureaus it had engaged. For instance, 4311 any errors in the original text were to be retained. Working from bound 4312 volumes or anything that could not be sheet-fed also constituted a factor 4313 eliminating companies that would have performed OCR. 4314 4315 AM requires 99.95 percent accuracy, which, though it sounds high, often 4316 means one or two errors per page. The initial batch of test samples 4317 contained several handwritten materials for which AM did not require 4318 text-coding. The results, ERWAY reported, were in all cases fairly 4319 comparable: for the most part, all three service bureaus achieved 99.95 4320 percent accuracy. AM was satisfied with the work but surprised at the cost. 4321 4322 As AM began converting whole collections, it retained the requirement for 4323 99.95 percent accuracy and added requirements for text-coding. AM needed 4324 to begin performing work more than three years ago before LC requirements 4325 for SGML applications had been established. Since AM's goal was simply 4326 to retain any of the intellectual content represented by the formatting 4327 of the document (which would be lost if one performed a straight ASCII 4328 conversion), AM used "SGML-like" codes. These codes resembled SGML tags 4329 but were used without the benefit of document-type definitions. AM found 4330 that many service bureaus were not yet SGML-proficient. 4331 4332 Additional factors influencing the approach AM took with respect to 4333 coding included: 1) the inability of any known microcomputer-based 4334 user-retrieval software to take advantage of SGML coding; and 2) the 4335 multiple inconsistencies in format of the older documents, which 4336 confirmed AM in its desire not to attempt to force the different formats 4337 to conform to a single document-type definition (DTD) and thus create the 4338 need for a separate DTD for each document. 4339 4340 The five text collections that AM has converted or is in the process of 4341 converting include a collection of eighteenth-century broadsides, a 4342 collection of pamphlets, two typescript document collections, and a 4343 collection of 150 books. 4344 4345 ERWAY next reviewed the results of AM's experience with rekeying, noting 4346 again that because the bulk of AM's materials are historical, the quality 4347 of the text often does not lend itself to OCR. While non-English 4348 speakers are less likely to guess or elaborate or correct typos in the 4349 original text, they are also less able to infer what we would; they also 4350 are nearly incapable of converting handwritten text. Another 4351 disadvantage of working with overseas keyers is that they are much less 4352 likely to telephone with questions, especially on the coding, with the 4353 result that they develop their own rules as they encounter new 4354 situations. 4355 4356 Government contracting procedures and time frames posed a major challenge 4357 to performing the conversion. Many service bureaus are not accustomed to 4358 retaining the image, even if they perform OCR. Thus, questions of image 4359 format and storage media were somewhat novel to many of them. ERWAY also 4360 remarked other problems in dealing with service bureaus, for example, 4361 their inability to perform text conversion from the kind of microfilm 4362 that LC uses for preservation purposes. 4363 4364 But quality control, in ERWAY's experience, was the most time-consuming 4365 aspect of contracting out conversion. AM has been attempting to perform 4366 a 10-percent quality review, looking at either every tenth document or 4367 every tenth page to make certain that the service bureaus are maintaining 4368 99.95 percent accuracy. But even if they are complying with the 4369 requirement for accuracy, finding errors produces a desire to correct 4370 them and, in turn, to clean up the whole collection, which defeats the 4371 purpose to some extent. Even a double entry requires a 4372 character-by-character comparison to the original to meet the accuracy 4373 requirement. LC is not accustomed to publish imperfect texts, which 4374 makes attempting to deal with the industry standard an emotionally 4375 fraught issue for AM. As was mentioned in the previous day's discussion, 4376 going from 99.95 to 99.99 percent accuracy usually doubles costs and 4377 means a third keying or another complete run-through of the text. 4378 4379 Although AM has learned much from its experiences with various collections 4380 and various service bureaus, ERWAY concluded pessimistically that no 4381 breakthrough has been achieved. Incremental improvements have occurred 4382 in some of the OCR technology, some of the processes, and some of the 4383 standards acceptances, which, though they may lead to somewhat lower costs, 4384 do not offer much encouragement to many people who are anxiously awaiting 4385 the day that the entire contents of LC are available on-line. 4386 4387 ****** 4388 4389 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 4390 ZIDAR * Several answers to why one attempts to perform full-text 4391 conversion * Per page cost of performing OCR * Typical problems 4392 encountered during editing * Editing poor copy OCR vs. rekeying * 4393 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 4394 4395 Judith ZIDAR, coordinator, National Agricultural Text Digitizing Program 4396 (NATDP), National Agricultural Library (NAL), offered several answers to 4397 the question of why one attempts to perform full-text conversion: 1) 4398 Text in an image can be read by a human but not by a computer, so of 4399 course it is not searchable and there is not much one can do with it. 2) 4400 Some material simply requires word-level access. For instance, the legal 4401 profession insists on full-text access to its material; with taxonomic or 4402 geographic material, which entails numerous names, one virtually requires 4403 word-level access. 3) Full text permits rapid browsing and searching, 4404 something that cannot be achieved in an image with today's technology. 4405 4) Text stored as ASCII and delivered in ASCII is standardized and highly 4406 portable. 5) People just want full-text searching, even those who do not 4407 know how to do it. NAL, for the most part, is performing OCR at an 4408 actual cost per average-size page of approximately $7. NAL scans the 4409 page to create the electronic image and passes it through the OCR device. 4410 4411 ZIDAR next rehearsed several typical problems encountered during editing. 4412 Praising the celerity of her student workers, ZIDAR observed that editing 4413 requires approximately five to ten minutes per page, assuming that there 4414 are no large tables to audit. Confusion among the three characters I, 1, 4415 and l, constitutes perhaps the most common problem encountered. Zeroes 4416 and O's also are frequently confused. Double M's create a particular 4417 problem, even on clean pages. They are so wide in most fonts that they 4418 touch, and the system simply cannot tell where one letter ends and the 4419 other begins. Complex page formats occasionally fail to columnate 4420 properly, which entails rescanning as though one were working with a 4421 single column, entering the ASCII, and decolumnating for better 4422 searching. With proportionally spaced text, OCR can have difficulty 4423 discerning what is a space and what are merely spaces between letters, as 4424 opposed to spaces between words, and therefore will merge text or break 4425 up words where it should not. 4426 4427 ZIDAR said that it can often take longer to edit a poor-copy OCR than to 4428 key it from scratch. NAL has also experimented with partial editing of 4429 text, whereby project workers go into and clean up the format, removing 4430 stray characters but not running a spell-check. NAL corrects typos in 4431 the title and authors' names, which provides a foothold for searching and 4432 browsing. Even extremely poor-quality OCR (e.g., 60-percent accuracy) 4433 can still be searched, because numerous words are correct, while the 4434 important words are probably repeated often enough that they are likely 4435 to be found correct somewhere. Librarians, however, cannot tolerate this 4436 situation, though end users seem more willing to use this text for 4437 searching, provided that NAL indicates that it is unedited. ZIDAR 4438 concluded that rekeying of text may be the best route to take, in spite 4439 of numerous problems with quality control and cost. 4440 4441 ****** 4442 4443 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 4444 DISCUSSION * Modifying an image before performing OCR * NAL's costs per 4445 page *AM's costs per page and experience with Federal Prison Industries * 4446 Elements comprising NATDP's costs per page * OCR and structured markup * 4447 Distinction between the structure of a document and its representation 4448 when put on the screen or printed * 4449 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 4450 4451 HOOTON prefaced the lengthy discussion that followed with several 4452 comments about modifying an image before one reaches the point of 4453 performing OCR. For example, in regard to an application containing a 4454 significant amount of redundant data, such as form-type data, numerous 4455 companies today are working on various kinds of form renewal, prior to 4456 going through a recognition process, by using dropout colors. Thus, 4457 acquiring access to form design or using electronic means are worth 4458 considering. HOOTON also noted that conversion usually makes or breaks 4459 one's imaging system. It is extremely important, extremely costly in 4460 terms of either capital investment or service, and determines the quality 4461 of the remainder of one's system, because it determines the character of 4462 the raw material used by the system. 4463 4464 Concerning the four projects undertaken by NAL, two inside and two 4465 performed by outside contractors, ZIDAR revealed that an in-house service 4466 bureau executed the first at a cost between $8 and $10 per page for 4467 everything, including building of the database. The project undertaken 4468 by the Consultative Group on International Agricultural Research (CGIAR) 4469 cost approximately $10 per page for the conversion, plus some expenses 4470 for the software and building of the database. The Acid Rain Project--a 4471 two-disk set produced by the University of Vermont, consisting of 4472 Canadian publications on acid rain--cost $6.70 per page for everything, 4473 including keying of the text, which was double keyed, scanning of the 4474 images, and building of the database. The in-house project offered 4475 considerable ease of convenience and greater control of the process. On 4476 the other hand, the service bureaus know their job and perform it 4477 expeditiously, because they have more people. 4478 4479 As a useful comparison, ERWAY revealed AM's costs as follows: $0.75 4480 cents to $0.85 cents per thousand characters, with an average page 4481 containing 2,700 characters. Requirements for coding and imaging 4482 increase the costs. Thus, conversion of the text, including the coding, 4483 costs approximately $3 per page. (This figure does not include the 4484 imaging and database-building included in the NAL costs.) AM also 4485 enjoyed a happy experience with Federal Prison Industries, which 4486 precluded the necessity of going through the request-for-proposal process 4487 to award a contract, because it is another government agency. The 4488 prisoners performed AM's rekeying just as well as other service bureaus 4489 and proved handy as well. AM shipped them the books, which they would 4490 photocopy on a book-edge scanner. They would perform the markup on 4491 photocopies, return the books as soon as they were done with them, 4492 perform the keying, and return the material to AM on WORM disks. 4493 4494 ZIDAR detailed the elements that constitute the previously noted cost of 4495 approximately $7 per page. Most significant is the editing, correction 4496 of errors, and spell-checkings, which though they may sound easy to 4497 perform require, in fact, a great deal of time. Reformatting text also 4498 takes a while, but a significant amount of NAL's expenses are for equipment, 4499 which was extremely expensive when purchased because it was one of the few 4500 systems on the market. The costs of equipment are being amortized over 4501 five years but are still quite high, nearly $2,000 per month. 4502 4503 HOCKEY raised a general question concerning OCR and the amount of editing 4504 required (substantial in her experience) to generate the kind of 4505 structured markup necessary for manipulating the text on the computer or 4506 loading it into any retrieval system. She wondered if the speakers could 4507 extend the previous question about the cost-benefit of adding or exerting 4508 structured markup. ERWAY noted that several OCR systems retain italics, 4509 bolding, and other spatial formatting. While the material may not be in 4510 the format desired, these systems possess the ability to remove the 4511 original materials quickly from the hands of the people performing the 4512 conversion, as well as to retain that information so that users can work 4513 with it. HOCKEY rejoined that the current thinking on markup is that one 4514 should not say that something is italic or bold so much as why it is that 4515 way. To be sure, one needs to know that something was italicized, but 4516 how can one get from one to the other? One can map from the structure to 4517 the typographic representation. 4518 4519 FLEISCHHAUER suggested that, given the 100 million items the Library 4520 holds, it may not be possible for LC to do more than report that a thing 4521 was in italics as opposed to why it was italics, although that may be 4522 desirable in some contexts. Promising to talk a bit during the afternoon 4523 session about several experiments OCLC performed on automatic recognition 4524 of document elements, and which they hoped to extend, WEIBEL said that in 4525 fact one can recognize the major elements of a document with a fairly 4526 high degree of reliability, at least as good as OCR. STEVENS drew a 4527 useful distinction between standard, generalized markup (i.e., defining 4528 for a document-type definition the structure of the document), and what 4529 he termed a style sheet, which had to do with italics, bolding, and other 4530 forms of emphasis. Thus, two different components are at work, one being 4531 the structure of the document itself (its logic), and the other being its 4532 representation when it is put on the screen or printed. 4533 4534 ****** 4535 4536 SESSION V. APPROACHES TO PREPARING ELECTRONIC TEXTS 4537 4538 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 4539 HOCKEY * Text in ASCII and the representation of electronic text versus 4540 an image * The need to look at ways of using markup to assist retrieval * 4541 The need for an encoding format that will be reusable and multifunctional 4542 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 4543 4544 Susan HOCKEY, director, Center for Electronic Texts in the Humanities 4545 (CETH), Rutgers and Princeton Universities, announced that one talk 4546 (WEIBEL's) was moved into this session from the morning and that David 4547 Packard was unable to attend. The session would attempt to focus more on 4548 what one can do with a text in ASCII and the representation of electronic 4549 text rather than just an image, what one can do with a computer that 4550 cannot be done with a book or an image. It would be argued that one can 4551 do much more than just read a text, and from that starting point one can 4552 use markup and methods of preparing the text to take full advantage of 4553 the capability of the computer. That would lead to a discussion of what 4554 the European Community calls REUSABILITY, what may better be termed 4555 DURABILITY, that is, how to prepare or make a text that will last a long 4556 time and that can be used for as many applications as possible, which 4557 would lead to issues of improving intellectual access. 4558 4559 HOCKEY urged the need to look at ways of using markup to facilitate retrieval, 4560 not just for referencing or to help locate an item that is retrieved, but also to put markup tags in 4561 a text to help retrieve the thing sought either with linguistic tagging or 4562 interpretation. HOCKEY also argued that little advancement had occurred in 4563 the software tools currently available for retrieving and searching text. 4564 She pressed the desideratum of going beyond Boolean searches and performing 4565 more sophisticated searching, which the insertion of more markup in the text 4566 would facilitate. Thinking about electronic texts as opposed to images means 4567 considering material that will never appear in print form, or print will not 4568 be its primary form, that is, material which only appears in electronic form. 4569 HOCKEY alluded to the history and the need for markup and tagging and 4570 electronic text, which was developed through the use of computers in the 4571 humanities; as MICHELSON had observed, Father Busa had started in 1949 4572 to prepare the first-ever text on the computer. 4573 4574 HOCKEY remarked several large projects, particularly in Europe, for the 4575 compilation of dictionaries, language studies, and language analysis, in 4576 which people have built up archives of text and have begun to recognize 4577 the need for an encoding format that will be reusable and multifunctional, 4578 that can be used not just to print the text, which may be assumed to be a 4579 byproduct of what one wants to do, but to structure it inside the computer 4580 so that it can be searched, built into a Hypertext system, etc. 4581 4582 ****** 4583 4584 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 4585 WEIBEL * OCLC's approach to preparing electronic text: retroconversion, 4586 keying of texts, more automated ways of developing data * Project ADAPT 4587 and the CORE Project * Intelligent character recognition does not exist * 4588 Advantages of SGML * Data should be free of procedural markup; 4589 descriptive markup strongly advocated * OCLC's interface illustrated * 4590 Storage requirements and costs for putting a lot of information on line * 4591 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 4592 4593 Stuart WEIBEL, senior research scientist, Online Computer Library Center, 4594 Inc. (OCLC), described OCLC's approach to preparing electronic text. He 4595 argued that the electronic world into which we are moving must 4596 accommodate not only the future but the past as well, and to some degree 4597 even the present. Thus, starting out at one end with retroconversion and 4598 keying of texts, one would like to move toward much more automated ways 4599 of developing data. 4600 4601 For example, Project ADAPT had to do with automatically converting 4602 document images into a structured document database with OCR text as 4603 indexing and also a little bit of automatic formatting and tagging of 4604 that text. The CORE project hosted by Cornell University, Bellcore, 4605 OCLC, the American Chemical Society, and Chemical Abstracts, constitutes 4606 WEIBEL's principal concern at the moment. This project is an example of 4607 converting text for which one already has a machine-readable version into 4608 a format more suitable for electronic delivery and database searching. 4609 (Since Michael LESK had previously described CORE, WEIBEL would say 4610 little concerning it.) Borrowing a chemical phrase, de novo synthesis, 4611 WEIBEL cited the Online Journal of Current Clinical Trials as an example 4612 of de novo electronic publishing, that is, a form in which the primary 4613 form of the information is electronic. 4614 4615 Project ADAPT, then, which OCLC completed a couple of years ago and in 4616 fact is about to resume, is a model in which one takes page images either 4617 in paper or microfilm and converts them automatically to a searchable 4618 electronic database, either on-line or local. The operating assumption 4619 is that accepting some blemishes in the data, especially for 4620 retroconversion of materials, will make it possible to accomplish more. 4621 Not enough money is available to support perfect conversion. 4622 4623 WEIBEL related several steps taken to perform image preprocessing 4624 (processing on the image before performing optical character 4625 recognition), as well as image postprocessing. He denied the existence 4626 of intelligent character recognition and asserted that what is wanted is 4627 page recognition, which is a long way off. OCLC has experimented with 4628 merging of multiple optical character recognition systems that will 4629 reduce errors from an unacceptable rate of 5 characters out of every 4630 l,000 to an unacceptable rate of 2 characters out of every l,000, but it 4631 is not good enough. It will never be perfect. 4632 4633 Concerning the CORE Project, WEIBEL observed that Bellcore is taking the 4634 topography files, extracting the page images, and converting those 4635 topography files to SGML markup. LESK hands that data off to OCLC, which 4636 builds that data into a Newton database, the same system that underlies 4637 the on-line system in virtually all of the reference products at OCLC. 4638 The long-term goal is to make the systems interoperable so that not just 4639 Bellcore's system and OCLC's system can access this data, but other 4640 systems can as well, and the key to that is the Z39.50 common command 4641 language and the full-text extension. Z39.50 is fine for MARC records, 4642 but is not enough to do it for full text (that is, make full texts 4643 interoperable). 4644 4645 WEIBEL next outlined the critical role of SGML for a variety of purposes, 4646 for example, as noted by HOCKEY, in the world of extremely large 4647 databases, using highly structured data to perform field searches. 4648 WEIBEL argued that by building the structure of the data in (i.e., the 4649 structure of the data originally on a printed page), it becomes easy to 4650 look at a journal article even if one cannot read the characters and know 4651 where the title or author is, or what the sections of that document would be. 4652 OCLC wants to make that structure explicit in the database, because it will 4653 be important for retrieval purposes. 4654 4655 The second big advantage of SGML is that it gives one the ability to 4656 build structure into the database that can be used for display purposes 4657 without contaminating the data with instructions about how to format 4658 things. The distinction lies between procedural markup, which tells one 4659 where to put dots on the page, and descriptive markup, which describes 4660 the elements of a document. 4661 4662 WEIBEL believes that there should be no procedural markup in the data at 4663 all, that the data should be completely unsullied by information about 4664 italics or boldness. That should be left up to the display device, 4665 whether that display device is a page printer or a screen display device. 4666 By keeping one's database free of that kind of contamination, one can 4667 make decisions down the road, for example, reorganize the data in ways 4668 that are not cramped by built-in notions of what should be italic and 4669 what should be bold. WEIBEL strongly advocated descriptive markup. As 4670 an example, he illustrated the index structure in the CORE data. With 4671 subsequent illustrated examples of markup, WEIBEL acknowledged the common 4672 complaint that SGML is hard to read in its native form, although markup 4673 decreases considerably once one gets into the body. Without the markup, 4674 however, one would not have the structure in the data. One can pass 4675 markup through a LaTeX processor and convert it relatively easily to a 4676 printed version of the document. 4677 4678 WEIBEL next illustrated an extremely cluttered screen dump of OCLC's 4679 system, in order to show as much as possible the inherent capability on 4680 the screen. (He noted parenthetically that he had become a supporter of 4681 X-Windows as a result of the progress of the CORE Project.) WEIBEL also 4682 illustrated the two major parts of the interface: l) a control box that 4683 allows one to generate lists of items, which resembles a small table of 4684 contents based on key words one wishes to search, and 2) a document 4685 viewer, which is a separate process in and of itself. He demonstrated 4686 how to follow links through the electronic database simply by selecting 4687 the appropriate button and bringing them up. He also noted problems that 4688 remain to be accommodated in the interface (e.g., as pointed out by LESK, 4689 what happens when users do not click on the icon for the figure). 4690 4691 Given the constraints of time, WEIBEL omitted a large number of ancillary 4692 items in order to say a few words concerning storage requirements and 4693 what will be required to put a lot of things on line. Since it is 4694 extremely expensive to reconvert all of this data, especially if it is 4695 just in paper form (and even if it is in electronic form in typesetting 4696 tapes), he advocated building journals electronically from the start. In 4697 that case, if one only has text graphics and indexing (which is all that 4698 one needs with de novo electronic publishing, because there is no need to 4699 go back and look at bit-maps of pages), one can get 10,000 journals of 4700 full text, or almost 6 million pages per year. These pages can be put in 4701 approximately 135 gigabytes of storage, which is not all that much, 4702 WEIBEL said. For twenty years, something less than three terabytes would 4703 be required. WEIBEL calculated the costs of storing this information as 4704 follows: If a gigabyte costs approximately $1,000, then a terabyte costs 4705 approximately $1 million to buy in terms of hardware. One also needs a 4706 building to put it in and a staff like OCLC to handle that information. 4707 So, to support a terabyte, multiply by five, which gives $5 million per 4708 year for a supported terabyte of data. 4709 4710 ****** 4711 4712 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 4713 DISCUSSION * Tapes saved by ACS are the typography files originally 4714 supporting publication of the journal * Cost of building tagged text into 4715 the database * 4716 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 4717 4718 During the question-and-answer period that followed WEIBEL's 4719 presentation, these clarifications emerged. The tapes saved by the 4720 American Chemical Society are the typography files that originally 4721 supported the publication of the journal. Although they are not tagged 4722 in SGML, they are tagged in very fine detail. Every single sentence is 4723 marked, all the registry numbers, all the publications issues, dates, and 4724 volumes. No cost figures on tagging material on a per-megabyte basis 4725 were available. Because ACS's typesetting system runs from tagged text, 4726 there is no extra cost per article. It was unknown what it costs ACS to 4727 keyboard the tagged text rather than just keyboard the text in the 4728 cheapest process. In other words, since one intends to publish things 4729 and will need to build tagged text into a typography system in any case, 4730 if one does that in such a way that it can drive not only typography but 4731 an electronic system (which is what ACS intends to do--move to SGML 4732 publishing), the marginal cost is zero. The marginal cost represents the 4733 cost of building tagged text into the database, which is small. 4734 4735 ****** 4736 4737 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 4738 SPERBERG-McQUEEN * Distinction between texts and computers * Implications 4739 of recognizing that all representation is encoding * Dealing with 4740 complicated representations of text entails the need for a grammar of 4741 documents * Variety of forms of formal grammars * Text as a bit-mapped 4742 image does not represent a serious attempt to represent text in 4743 electronic form * SGML, the TEI, document-type declarations, and the 4744 reusability and longevity of data * TEI conformance explicitly allows 4745 extension or modification of the TEI tag set * Administrative background 4746 of the TEI * Several design goals for the TEI tag set * An absolutely 4747 fixed requirement of the TEI Guidelines * Challenges the TEI has 4748 attempted to face * Good texts not beyond economic feasibility * The 4749 issue of reproducibility or processability * The issue of mages as 4750 simulacra for the text redux * One's model of text determines what one's 4751 software can do with a text and has economic consequences * 4752 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 4753 4754 Prior to speaking about SGML and markup, Michael SPERBERG-McQUEEN, editor, 4755 Text Encoding Initiative (TEI), University of Illinois-Chicago, first drew 4756 a distinction between texts and computers: Texts are abstract cultural 4757 and linguistic objects while computers are complicated physical devices, 4758 he said. Abstract objects cannot be placed inside physical devices; with 4759 computers one can only represent text and act upon those representations. 4760 4761 The recognition that all representation is encoding, SPERBERG-McQUEEN 4762 argued, leads to the recognition of two things: 1) The topic description 4763 for this session is slightly misleading, because there can be no discussion 4764 of pros and cons of text-coding unless what one means is pros and cons of 4765 working with text with computers. 2) No text can be represented in a 4766 computer without some sort of encoding; images are one way of encoding text, 4767 ASCII is another, SGML yet another. There is no encoding without some 4768 information loss, that is, there is no perfect reproduction of a text that 4769 allows one to do away with the original. Thus, the question becomes, 4770 What is the most useful representation of text for a serious work? 4771 This depends on what kind of serious work one is talking about. 4772 4773 The projects demonstrated the previous day all involved highly complex 4774 information and fairly complex manipulation of the textual material. 4775 In order to use that complicated information, one has to calculate it 4776 slowly or manually and store the result. It needs to be stored, therefore, 4777 as part of one's representation of the text. Thus, one needs to store the 4778 structure in the text. To deal with complicated representations of text, 4779 one needs somehow to control the complexity of the representation of a text; 4780 that means one needs a way of finding out whether a document and an 4781 electronic representation of a document is legal or not; and that 4782 means one needs a grammar of documents. 4783 4784 SPERBERG-McQUEEN discussed the variety of forms of formal grammars, 4785 implicit and explicit, as applied to text, and their capabilities. He 4786 argued that these grammars correspond to different models of text that 4787 different developers have. For example, one implicit model of the text 4788 is that there is no internal structure, but just one thing after another, 4789 a few characters and then perhaps a start-title command, and then a few 4790 more characters and an end-title command. SPERBERG-McQUEEN also 4791 distinguished several kinds of text that have a sort of hierarchical 4792 structure that is not very well defined, which, typically, corresponds 4793 to grammars that are not very well defined, as well as hierarchies that 4794 are very well defined (e.g., the Thesaurus Linguae Graecae) and extremely 4795 complicated things such as SGML, which handle strictly hierarchical data 4796 very nicely. 4797 4798 SPERBERG-McQUEEN conceded that one other model not illustrated on his two 4799 displays was the model of text as a bit-mapped image, an image of a page, 4800 and confessed to having been converted to a limited extent by the 4801 Workshop to the view that electronic images constitute a promising, 4802 probably superior alternative to microfilming. But he was not convinced 4803 that electronic images represent a serious attempt to represent text in 4804 electronic form. Many of their problems stem from the fact that they are 4805 not direct attempts to represent the text but attempts to represent the 4806 page, thus making them representations of representations. 4807 4808 In this situation of increasingly complicated textual information and the 4809 need to control that complexity in a useful way (which begs the question 4810 of the need for good textual grammars), one has the introduction of SGML. 4811 With SGML, one can develop specific document-type declarations 4812 for specific text types or, as with the TEI, attempts to generate 4813 general document-type declarations that can handle all sorts of text. 4814 The TEI is an attempt to develop formats for text representation that 4815 will ensure the kind of reusability and longevity of data discussed earlier. 4816 It offers a way to stay alive in the state of permanent technological 4817 revolution. 4818 4819 It has been a continuing challenge in the TEI to create document grammars 4820 that do some work in controlling the complexity of the textual object but 4821 also allowing one to represent the real text that one will find. 4822 Fundamental to the notion of the TEI is that TEI conformance allows one 4823 the ability to extend or modify the TEI tag set so that it fits the text 4824 that one is attempting to represent. 4825 4826 SPERBERG-McQUEEN next outlined the administrative background of the TEI. 4827 The TEI is an international project to develop and disseminate guidelines 4828 for the encoding and interchange of machine-readable text. It is 4829 sponsored by the Association for Computers in the Humanities, the 4830 Association for Computational Linguistics, and the Association for 4831 Literary and Linguistic Computing. Representatives of numerous other 4832 professional societies sit on its advisory board. The TEI has a number 4833 of affiliated projects that have provided assistance by testing drafts of 4834 the guidelines. 4835 4836 Among the design goals for the TEI tag set, the scheme first of all must 4837 meet the needs of research, because the TEI came out of the research 4838 community, which did not feel adequately served by existing tag sets. 4839 The tag set must be extensive as well as compatible with existing and 4840 emerging standards. In 1990, version 1.0 of the Guidelines was released 4841 (SPERBERG-McQUEEN illustrated their contents). 4842 4843 SPERBERG-McQUEEN noted that one problem besetting electronic text has 4844 been the lack of adequate internal or external documentation for many 4845 existing electronic texts. The TEI guidelines as currently formulated 4846 contain few fixed requirements, but one of them is this: There must 4847 always be a document header, an in-file SGML tag that provides 4848 1) a bibliographic description of the electronic object one is talking 4849 about (that is, who included it, when, what for, and under which title); 4850 and 2) the copy text from which it was derived, if any. If there was 4851 no copy text or if the copy text is unknown, then one states as much. 4852 Version 2.0 of the Guidelines was scheduled to be completed in fall 1992 4853 and a revised third version is to be presented to the TEI advisory board 4854 for its endorsement this coming winter. The TEI itself exists to provide 4855 a markup language, not a marked-up text. 4856 4857 Among the challenges the TEI has attempted to face is the need for a 4858 markup language that will work for existing projects, that is, handle the 4859 level of markup that people are using now to tag only chapter, section, 4860 and paragraph divisions and not much else. At the same time, such a 4861 language also will be able to scale up gracefully to handle the highly 4862 detailed markup which many people foresee as the future destination of 4863 much electronic text, and which is not the future destination but the 4864 present home of numerous electronic texts in specialized areas. 4865 4866 SPERBERG-McQUEEN dismissed the lowest-common-denominator approach as 4867 unable to support the kind of applications that draw people who have 4868 never been in the public library regularly before, and make them come 4869 back. He advocated more interesting text and more intelligent text. 4870 Asserting that it is not beyond economic feasibility to have good texts, 4871 SPERBERG-McQUEEN noted that the TEI Guidelines listing 200-odd tags 4872 contains tags that one is expected to enter every time the relevant 4873 textual feature occurs. It contains all the tags that people need now, 4874 and it is not expected that everyone will tag things in the same way. 4875 4876 The question of how people will tag the text is in large part a function 4877 of their reaction to what SPERBERG-McQUEEN termed the issue of 4878 reproducibility. What one needs to be able to reproduce are the things 4879 one wants to work with. Perhaps a more useful concept than that of 4880 reproducibility or recoverability is that of processability, that is, 4881 what can one get from an electronic text without reading it again 4882 in the original. He illustrated this contention with a page from 4883 Jan Comenius's bilingual Introduction to Latin. 4884 4885 SPERBERG-McQUEEN returned at length to the issue of images as simulacra 4886 for the text, in order to reiterate his belief that in the long run more 4887 than images of pages of particular editions of the text are needed, 4888 because just as second-generation photocopies and second-generation 4889 microfilm degenerate, so second-generation representations tend to 4890 degenerate, and one tends to overstress some relatively trivial aspects 4891 of the text such as its layout on the page, which is not always 4892 significant, despite what the text critics might say, and slight other 4893 pieces of information such as the very important lexical ties between the 4894 English and Latin versions of Comenius's bilingual text, for example. 4895 Moreover, in many crucial respects it is easy to fool oneself concerning 4896 what a scanned image of the text will accomplish. For example, in order 4897 to study the transmission of texts, information concerning the text 4898 carrier is necessary, which scanned images simply do not always handle. 4899 Further, even the high-quality materials being produced at Cornell use 4900 much of the information that one would need if studying those books as 4901 physical objects. It is a choice that has been made. It is an arguably 4902 justifiable choice, but one does not know what color those pen strokes in 4903 the margin are or whether there was a stain on the page, because it has 4904 been filtered out. One does not know whether there were rips in the page 4905 because they do not show up, and on a couple of the marginal marks one 4906 loses half of the mark because the pen is very light and the scanner 4907 failed to pick it up, and so what is clearly a checkmark in the margin of 4908 the original becomes a little scoop in the margin of the facsimile. 4909 Standard problems for facsimile editions, not new to electronics, but 4910 also true of light-lens photography, and are remarked here because it is 4911 important that we not fool ourselves that even if we produce a very nice 4912 image of this page with good contrast, we are not replacing the 4913 manuscript any more than microfilm has replaced the manuscript. 4914 4915 The TEI comes from the research community, where its first allegiance 4916 lies, but it is not just an academic exercise. It has relevance far 4917 beyond those who spend all of their time studying text, because one's 4918 model of text determines what one's software can do with a text. Good 4919 models lead to good software. Bad models lead to bad software. That has 4920 economic consequences, and it is these economic consequences that have 4921 led the European Community to help support the TEI, and that will lead, 4922 SPERBERG-McQUEEN hoped, some software vendors to realize that if they 4923 provide software with a better model of the text they can make a killing. 4924 4925 ****** 4926 4927 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 4928 DISCUSSION * Implications of different DTDs and tag sets * ODA versus SGML * 4929 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 4930 4931 During the discussion that followed, several additional points were made. 4932 Neither AAP (i.e., Association of American Publishers) nor CALS (i.e., 4933 Computer-aided Acquisition and Logistics Support) has a document-type 4934 definition for ancient Greek drama, although the TEI will be able to 4935 handle that. Given this state of affairs and assuming that the 4936 technical-journal producers and the commercial vendors decide to use the 4937 other two types, then an institution like the Library of Congress, which 4938 might receive all of their publications, would have to be able to handle 4939 three different types of document definitions and tag sets and be able to 4940 distinguish among them. 4941 4942 Office Document Architecture (ODA) has some advantages that flow from its 4943 tight focus on office documents and clear directions for implementation. 4944 Much of the ODA standard is easier to read and clearer at first reading 4945 than the SGML standard, which is extremely general. What that means is 4946 that if one wants to use graphics in TIFF and ODA, one is stuck, because 4947 ODA defines graphics formats while TIFF does not, whereas SGML says the 4948 world is not waiting for this work group to create another graphics format. 4949 What is needed is an ability to use whatever graphics format one wants. 4950 4951 The TEI provides a socket that allows one to connect the SGML document to 4952 the graphics. The notation that the graphics are in is clearly a choice 4953 that one needs to make based on her or his environment, and that is one 4954 advantage. SGML is less megalomaniacal in attempting to define formats 4955 for all kinds of information, though more megalomaniacal in attempting to 4956 cover all sorts of documents. The other advantage is that the model of 4957 text represented by SGML is simply an order of magnitude richer and more 4958 flexible than the model of text offered by ODA. Both offer hierarchical 4959 structures, but SGML recognizes that the hierarchical model of the text 4960 that one is looking at may not have been in the minds of the designers, 4961 whereas ODA does not. 4962 4963 ODA is not really aiming for the kind of document that the TEI wants to 4964 encompass. The TEI can handle the kind of material ODA has, as well as a 4965 significantly broader range of material. ODA seems to be very much 4966 focused on office documents, which is what it started out being called-- 4967 office document architecture. 4968 4969 ****** 4970 4971 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 4972 CALALUCA * Text-encoding from a publisher's perspective * 4973 Responsibilities of a publisher * Reproduction of Migne's Latin series 4974 whole and complete with SGML tags based on perceived need and expected 4975 use * Particular decisions arising from the general decision to produce 4976 and publish PLD * 4977 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 4978 4979 The final speaker in this session, Eric CALALUCA, vice president, 4980 Chadwyck-Healey, Inc., spoke from the perspective of a publisher re 4981 text-encoding, rather than as one qualified to discuss methods of 4982 encoding data, and observed that the presenters sitting in the room, 4983 whether they had chosen to or not, were acting as publishers: making 4984 choices, gathering data, gathering information, and making assessments. 4985 CALALUCA offered the hard-won conviction that in publishing very large 4986 text files (such as PLD), one cannot avoid making personal judgments of 4987 appropriateness and structure. 4988 4989 In CALALUCA's view, encoding decisions stem from prior judgments. Two 4990 notions have become axioms for him in the consideration of future sources 4991 for electronic publication: 1) electronic text publishing is as personal 4992 as any other kind of publishing, and questions of if and how to encode 4993 the data are simply a consequence of that prior decision; 2) all 4994 personal decisions are open to criticism, which is unavoidable. 4995 4996 CALALUCA rehearsed his role as a publisher or, better, as an intermediary 4997 between what is viewed as a sound idea and the people who would make use 4998 of it. Finding the specialist to advise in this process is the core of 4999 that function. The publisher must monitor and hug the fine line between 5000 giving users what they want and suggesting what they might need. One 5001 responsibility of a publisher is to represent the desires of scholars and 5002 research librarians as opposed to bullheadedly forcing them into areas 5003 they would not choose to enter. 5004 5005 CALALUCA likened the questions being raised today about data structure 5006 and standards to the decisions faced by the Abbe Migne himself during 5007 production of the Patrologia series in the mid-nineteenth century. 5008 Chadwyck-Healey's decision to reproduce Migne's Latin series whole and 5009 complete with SGML tags was also based upon a perceived need and an 5010 expected use. In the same way that Migne's work came to be far more than 5011 a simple handbook for clerics, PLD is already far more than a database 5012 for theologians. It is a bedrock source for the study of Western 5013 civilization, CALALUCA asserted. 5014 5015 In regard to the decision to produce and publish PLD, the editorial board 5016 offered direct judgments on the question of appropriateness of these 5017 texts for conversion, their encoding and their distribution, and 5018 concluded that the best possible project was one that avoided overt 5019 intrusions or exclusions in so important a resource. Thus, the general 5020 decision to transmit the original collection as clearly as possible with 5021 the widest possible avenues for use led to other decisions: 1) To encode 5022 the data or not, SGML or not, TEI or not. Again, the expected user 5023 community asserted the need for normative tagging structures of important 5024 humanities texts, and the TEI seemed the most appropriate structure for 5025 that purpose. Research librarians, who are trained to view the larger 5026 impact of electronic text sources on 80 or 90 or 100 doctoral 5027 disciplines, loudly approved the decision to include tagging. They see 5028 what is coming better than the specialist who is completely focused on 5029 one edition of Ambrose's De Anima, and they also understand that the 5030 potential uses exceed present expectations. 2) What will be tagged and 5031 what will not. Once again, the board realized that one must tag the 5032 obvious. But in no way should one attempt to identify through encoding 5033 schemes every single discrete area of a text that might someday be 5034 searched. That was another decision. Searching by a column number, an 5035 author, a word, a volume, permitting combination searches, and tagging 5036 notations seemed logical choices as core elements. 3) How does one make 5037 the data available? Tieing it to a CD-ROM edition creates limitations, 5038 but a magnetic tape file that is very large, is accompanied by the 5039 encoding specifications, and that allows one to make local modifications 5040 also allows one to incorporate any changes one may desire within the 5041 bounds of private research, though exporting tag files from a CD-ROM 5042 could serve just as well. Since no one on the board could possibly 5043 anticipate each and every way in which a scholar might choose to mine 5044 this data bank, it was decided to satisfy the basics and make some 5045 provisions for what might come. 4) Not to encode the database would rob 5046 it of the interchangeability and portability these important texts should 5047 accommodate. For CALALUCA, the extensive options presented by full-text 5048 searching require care in text selection and strongly support encoding of 5049 data to facilitate the widest possible search strategies. Better 5050 software can always be created, but summoning the resources, the people, 5051 and the energy to reconvert the text is another matter. 5052 5053 PLD is being encoded, captured, and distributed, because to 5054 Chadwyck-Healey and the board it offers the widest possible array of 5055 future research applications that can be seen today. CALALUCA concluded 5056 by urging the encoding of all important text sources in whatever way 5057 seems most appropriate and durable at the time, without blanching at the 5058 thought that one's work may require emendation in the future. (Thus, 5059 Chadwyck-Healey produced a very large humanities text database before the 5060 final release of the TEI Guidelines.) 5061 5062 ****** 5063 5064 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 5065 DISCUSSION * Creating texts with markup advocated * Trends in encoding * 5066 The TEI and the issue of interchangeability of standards * A 5067 misconception concerning the TEI * Implications for an institution like 5068 LC in the event that a multiplicity of DTDs develops * Producing images 5069 as a first step towards possible conversion to full text through 5070 character recognition * The AAP tag sets as a common starting point and 5071 the need for caution * 5072 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 5073 5074 HOCKEY prefaced the discussion that followed with several comments in 5075 favor of creating texts with markup and on trends in encoding. In the 5076 future, when many more texts are available for on-line searching, real 5077 problems in finding what is wanted will develop, if one is faced with 5078 millions of words of data. It therefore becomes important to consider 5079 putting markup in texts to help searchers home in on the actual things 5080 they wish to retrieve. Various approaches to refining retrieval methods 5081 toward this end include building on a computer version of a dictionary 5082 and letting the computer look up words in it to obtain more information 5083 about the semantic structure or semantic field of a word, its grammatical 5084 structure, and syntactic structure. 5085 5086 HOCKEY commented on the present keen interest in the encoding world 5087 in creating: 1) machine-readable versions of dictionaries that can be 5088 initially tagged in SGML, which gives a structure to the dictionary entry; 5089 these entries can then be converted into a more rigid or otherwise 5090 different database structure inside the computer, which can be treated as 5091 a dynamic tool for searching mechanisms; 2) large bodies of text to study 5092 the language. In order to incorporate more sophisticated mechanisms, 5093 more about how words behave needs to be known, which can be learned in 5094 part from information in dictionaries. However, the last ten years have 5095 seen much interest in studying the structure of printed dictionaries 5096 converted into computer-readable form. The information one derives about 5097 many words from those is only partial, one or two definitions of the 5098 common or the usual meaning of a word, and then numerous definitions of 5099 unusual usages. If the computer is using a dictionary to help retrieve 5100 words in a text, it needs much more information about the common usages, 5101 because those are the ones that occur over and over again. Hence the 5102 current interest in developing large bodies of text in computer-readable 5103 form in order to study the language. Several projects are engaged in 5104 compiling, for example, 100 million words. HOCKEY described one with 5105 which she was associated briefly at Oxford University involving 5106 compilation of 100 million words of British English: about 10 percent of 5107 that will contain detailed linguistic tagging encoded in SGML; it will 5108 have word class taggings, with words identified as nouns, verbs, 5109 adjectives, or other parts of speech. This tagging can then be used by 5110 programs which will begin to learn a bit more about the structure of the 5111 language, and then, can go to tag more text. 5112 5113 HOCKEY said that the more that is tagged accurately, the more one can 5114 refine the tagging process and thus the bigger body of text one can build 5115 up with linguistic tagging incorporated into it. Hence, the more tagging 5116 or annotation there is in the text, the more one may begin to learn about 5117 language and the more it will help accomplish more intelligent OCR. She 5118 recommended the development of software tools that will help one begin to 5119 understand more about a text, which can then be applied to scanning 5120 images of that text in that format and to using more intelligence to help 5121 one interpret or understand the text. 5122 5123 HOCKEY posited the need to think about common methods of text-encoding 5124 for a long time to come, because building these large bodies of text is 5125 extremely expensive and will only be done once. 5126 5127 In the more general discussion on approaches to encoding that followed, 5128 these points were made: 5129 5130 BESSER identified the underlying problem with standards that all have to 5131 struggle with in adopting a standard, namely, the tension between a very 5132 highly defined standard that is very interchangeable but does not work 5133 for everyone because something is lacking, and a standard that is less 5134 defined, more open, more adaptable, but less interchangeable. Contending 5135 that the way in which people use SGML is not sufficiently defined, BESSER 5136 wondered 1) if people resist the TEI because they think it is too defined 5137 in certain things they do not fit into, and 2) how progress with 5138 interchangeability can be made without frightening people away. 5139 5140 SPERBERG-McQUEEN replied that the published drafts of the TEI had met 5141 with surprisingly little objection on the grounds that they do not allow 5142 one to handle X or Y or Z. Particular concerns of the affiliated 5143 projects have led, in practice, to discussions of how extensions are to 5144 be made; the primary concern of any project has to be how it can be 5145 represented locally, thus making interchange secondary. The TEI has 5146 received much criticism based on the notion that everything in it is 5147 required or even recommended, which, as it happens, is a misconception 5148 from the beginning, because none of it is required and very little is 5149 actually actively recommended for all cases, except that one document 5150 one's source. 5151 5152 SPERBERG-McQUEEN agreed with BESSER about this trade-off: all the 5153 projects in a set of twenty TEI-conformant projects will not necessarily 5154 tag the material in the same way. One result of the TEI will be that the 5155 easiest problems will be solved--those dealing with the external form of 5156 the information; but the problem that is hardest in interchange is that 5157 one is not encoding what another wants, and vice versa. Thus, after 5158 the adoption of a common notation, the differences in the underlying 5159 conceptions of what is interesting about texts become more visible. 5160 The success of a standard like the TEI will lie in the ability of 5161 the recipient of interchanged texts to use some of what it contains 5162 and to add the information that was not encoded that one wants, in a 5163 layered way, so that texts can be gradually enriched and one does not 5164 have to put in everything all at once. Hence, having a well-behaved 5165 markup scheme is important. 5166 5167 STEVENS followed up on the paradoxical analogy that BESSER alluded to in 5168 the example of the MARC records, namely, the formats that are the same 5169 except that they are different. STEVENS drew a parallel between 5170 document-type definitions and MARC records for books and serials and maps, 5171 where one has a tagging structure and there is a text-interchange. 5172 STEVENS opined that the producers of the information will set the terms 5173 for the standard (i.e., develop document-type definitions for the users 5174 of their products), creating a situation that will be problematical for 5175 an institution like the Library of Congress, which will have to deal with 5176 the DTDs in the event that a multiplicity of them develops. Thus, 5177 numerous people are seeking a standard but cannot find the tag set that 5178 will be acceptable to them and their clients. SPERBERG-McQUEEN agreed 5179 with this view, and said that the situation was in a way worse: attempting 5180 to unify arbitrary DTDs resembled attempting to unify a MARC record with a 5181 bibliographic record done according to the Prussian instructions. 5182 According to STEVENS, this situation occurred very early in the process. 5183 5184 WATERS recalled from early discussions on Project Open Book the concern 5185 of many people that merely by producing images, POB was not really 5186 enhancing intellectual access to the material. Nevertheless, not wishing 5187 to overemphasize the opposition between imaging and full text, WATERS 5188 stated that POB views getting the images as a first step toward possibly 5189 converting to full text through character recognition, if the technology 5190 is appropriate. WATERS also emphasized that encoding is involved even 5191 with a set of images. 5192 5193 SPERBERG-McQUEEN agreed with WATERS that one can create an SGML document 5194 consisting wholly of images. At first sight, organizing graphic images 5195 with an SGML document may not seem to offer great advantages, but the 5196 advantages of the scheme WATERS described would be precisely that 5197 ability to move into something that is more of a multimedia document: 5198 a combination of transcribed text and page images. WEIBEL concurred in 5199 this judgment, offering evidence from Project ADAPT, where a page is 5200 divided into text elements and graphic elements, and in fact the text 5201 elements are organized by columns and lines. These lines may be used as 5202 the basis for distributing documents in a network environment. As one 5203 develops software intelligent enough to recognize what those elements 5204 are, it makes sense to apply SGML to an image initially, that may, in 5205 fact, ultimately become more and more text, either through OCR or edited 5206 OCR or even just through keying. For WATERS, the labor of composing the 5207 document and saying this set of documents or this set of images belongs 5208 to this document constitutes a significant investment. 5209 5210 WEIBEL also made the point that the AAP tag sets, while not excessively 5211 prescriptive, offer a common starting point; they do not define the 5212 structure of the documents, though. They have some recommendations about 5213 DTDs one could use as examples, but they do just suggest tag sets. For 5214 example, the CORE project attempts to use the AAP markup as much as 5215 possible, but there are clearly areas where structure must be added. 5216 That in no way contradicts the use of AAP tag sets. 5217 5218 SPERBERG-McQUEEN noted that the TEI prepared a long working paper early 5219 on about the AAP tag set and what it lacked that the TEI thought it 5220 needed, and a fairly long critique of the naming conventions, which has 5221 led to a very different style of naming in the TEI. He stressed the 5222 importance of the opposition between prescriptive markup, the kind that a 5223 publisher or anybody can do when producing documents de novo, and 5224 descriptive markup, in which one has to take what the text carrier 5225 provides. In these particular tag sets it is easy to overemphasize this 5226 opposition, because the AAP tag set is extremely flexible. Even if one 5227 just used the DTDs, they allow almost anything to appear almost anywhere. 5228 5229 ****** 5230 5231 SESSION VI. COPYRIGHT ISSUES 5232 5233 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 5234 PETERS * Several cautions concerning copyright in an electronic 5235 environment * Review of copyright law in the United States * The notion 5236 of the public good and the desirability of incentives to promote it * 5237 What copyright protects * Works not protected by copyright * The rights 5238 of copyright holders * Publishers' concerns in today's electronic 5239 environment * Compulsory licenses * The price of copyright in a digital 5240 medium and the need for cooperation * Additional clarifications * Rough 5241 justice oftentimes the outcome in numerous copyright matters * Copyright 5242 in an electronic society * Copyright law always only sets up the 5243 boundaries; anything can be changed by contract * 5244 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 5245 5246 Marybeth PETERS, policy planning adviser to the Register of Copyrights, 5247 Library of Congress, made several general comments and then opened the 5248 floor to discussion of subjects of interest to the audience. 5249 5250 Having attended several sessions in an effort to gain a sense of what 5251 people did and where copyright would affect their lives, PETERS expressed 5252 the following cautions: 5253 5254 * If one takes and converts materials and puts them in new forms, 5255 then, from a copyright point of view, one is creating something and 5256 will receive some rights. 5257 5258 * However, if what one is converting already exists, a question 5259 immediately arises about the status of the materials in question. 5260 5261 * Putting something in the public domain in the United States offers 5262 some freedom from anxiety, but distributing it throughout the world 5263 on a network is another matter, even if one has put it in the public 5264 domain in the United States. Re foreign laws, very frequently a 5265 work can be in the public domain in the United States but protected 5266 in other countries. Thus, one must consider all of the places a 5267 work may reach, lest one unwittingly become liable to being faced 5268 with a suit for copyright infringement, or at least a letter 5269 demanding discussion of what one is doing. 5270 5271 PETERS reviewed copyright law in the United States. The U.S. 5272 Constitution effectively states that Congress has the power to enact 5273 copyright laws for two purposes: 1) to encourage the creation and 5274 dissemination of intellectual works for the good of society as a whole; 5275 and, significantly, 2) to give creators and those who package and 5276 disseminate materials the economic rewards that are due them. 5277 5278 Congress strives to strike a balance, which at times can become an 5279 emotional issue. The United States has never accepted the notion of the 5280 natural right of an author so much as it has accepted the notion of the 5281 public good and the desirability of incentives to promote it. This state 5282 of affairs, however, has created strains on the international level and 5283 is the reason for several of the differences in the laws that we have. 5284 Today the United States protects almost every kind of work that can be 5285 called an expression of an author. The standard for gaining copyright 5286 protection is simply originality. This is a low standard and means that 5287 a work is not copied from something else, as well as shows a certain 5288 minimal amount of authorship. One can also acquire copyright protection 5289 for making a new version of preexisting material, provided it manifests 5290 some spark of creativity. 5291 5292 However, copyright does not protect ideas, methods, systems--only the way 5293 that one expresses those things. Nor does copyright protect anything 5294 that is mechanical, anything that does not involve choice, or criteria 5295 concerning whether or not one should do a thing. For example, the 5296 results of a process called declicking, in which one mechanically removes 5297 impure sounds from old recordings, are not copyrightable. On the other 5298 hand, the choice to record a song digitally and to increase the sound of 5299 violins or to bring up the tympani constitutes the results of conversion 5300 that are copyrightable. Moreover, if a work is protected by copyright in 5301 the United States, one generally needs the permission of the copyright 5302 owner to convert it. Normally, who will own the new--that is, converted- 5303 -material is a matter of contract. In the absence of a contract, the 5304 person who creates the new material is the author and owner. But people 5305 do not generally think about the copyright implications until after the 5306 fact. PETERS stressed the need when dealing with copyrighted works to 5307 think about copyright in advance. One's bargaining power is much greater 5308 up front than it is down the road. 5309 5310 PETERS next discussed works not protected by copyright, for example, any 5311 work done by a federal employee as part of his or her official duties is 5312 in the public domain in the United States. The issue is not wholly free 5313 of doubt concerning whether or not the work is in the public domain 5314 outside the United States. Other materials in the public domain include: 5315 any works published more than seventy-five years ago, and any work 5316 published in the United States more than twenty-eight years ago, whose 5317 copyright was not renewed. In talking about the new technology and 5318 putting material in a digital form to send all over the world, PETERS 5319 cautioned, one must keep in mind that while the rights may not be an 5320 issue in the United States, they may be in different parts of the world, 5321 where most countries previously employed a copyright term of the life of 5322 the author plus fifty years. 5323 5324 PETERS next reviewed the economics of copyright holding. Simply, 5325 economic rights are the rights to control the reproduction of a work in 5326 any form. They belong to the author, or in the case of a work made for 5327 hire, the employer. The second right, which is critical to conversion, 5328 is the right to change a work. The right to make new versions is perhaps 5329 one of the most significant rights of authors, particularly in an 5330 electronic world. The third right is the right to publish the work and 5331 the right to disseminate it, something that everyone who deals in an 5332 electronic medium needs to know. The basic rule is if a copy is sold, 5333 all rights of distribution are extinguished with the sale of that copy. 5334 The key is that it must be sold. A number of companies overcome this 5335 obstacle by leasing or renting their product. These companies argue that 5336 if the material is rented or leased and not sold, they control the uses 5337 of a work. The fourth right, and one very important in a digital world, 5338 is a right of public performance, which means the right to show the work 5339 sequentially. For example, copyright owners control the showing of a 5340 CD-ROM product in a public place such as a public library. The reverse 5341 side of public performance is something called the right of public 5342 display. Moral rights also exist, which at the federal level apply only 5343 to very limited visual works of art, but in theory may apply under 5344 contract and other principles. Moral rights may include the right of an 5345 author to have his or her name on a work, the right of attribution, and 5346 the right to object to distortion or mutilation--the right of integrity. 5347 5348 The way copyright law is worded gives much latitude to activities such as 5349 preservation; to use of material for scholarly and research purposes when 5350 the user does not make multiple copies; and to the generation of 5351 facsimile copies of unpublished works by libraries for themselves and 5352 other libraries. But the law does not allow anyone to become the 5353 distributor of the product for the entire world. In today's electronic 5354 environment, publishers are extremely concerned that the entire world is 5355 networked and can obtain the information desired from a single copy in a 5356 single library. Hence, if there is to be only one sale, which publishers 5357 may choose to live with, they will obtain their money in other ways, for 5358 example, from access and use. Hence, the development of site licenses 5359 and other kinds of agreements to cover what publishers believe they 5360 should be compensated for. Any solution that the United States takes 5361 today has to consider the international arena. 5362 5363 Noting that the United States is a member of the Berne Convention and 5364 subscribes to its provisions, PETERS described the permissions process. 5365 She also defined compulsory licenses. A compulsory license, of which the 5366 United States has had a few, builds into the law the right to use a work 5367 subject to certain terms and conditions. In the international arena, 5368 however, the ability to use compulsory licenses is extremely limited. 5369 Thus, clearinghouses and other collectives comprise one option that has 5370 succeeded in providing for use of a work. Often overlooked when one 5371 begins to use copyrighted material and put products together is how 5372 expensive the permissions process and managing it is. According to 5373 PETERS, the price of copyright in a digital medium, whatever solution is 5374 worked out, will include managing and assembling the database. She 5375 strongly recommended that publishers and librarians or people with 5376 various backgrounds cooperate to work out administratively feasible 5377 systems, in order to produce better results. 5378 5379 In the lengthy question-and-answer period that followed PETERS's 5380 presentation, the following points emerged: 5381 5382 * The Copyright Office maintains that anything mechanical and 5383 totally exhaustive probably is not protected. In the event that 5384 what an individual did in developing potentially copyrightable 5385 material is not understood, the Copyright Office will ask about the 5386 creative choices the applicant chose to make or not to make. As a 5387 practical matter, if one believes she or he has made enough of those 5388 choices, that person has a right to assert a copyright and someone 5389 else must assert that the work is not copyrightable. The more 5390 mechanical, the more automatic, a thing is, the less likely it is to 5391 be copyrightable. 5392 5393 * Nearly all photographs are deemed to be copyrightable, but no one 5394 worries about them much, because everyone is free to take the same 5395 image. Thus, a photographic copyright represents what is called a 5396 "thin" copyright. The photograph itself must be duplicated, in 5397 order for copyright to be violated. 5398 5399 * The Copyright Office takes the position that X-rays are not 5400 copyrightable because they are mechanical. It can be argued 5401 whether or not image enhancement in scanning can be protected. One 5402 must exercise care with material created with public funds and 5403 generally in the public domain. An article written by a federal 5404 employee, if written as part of official duties, is not 5405 copyrightable. However, control over a scientific article written 5406 by a National Institutes of Health grantee (i.e., someone who 5407 receives money from the U.S. government), depends on NIH policy. If 5408 the government agency has no policy (and that policy can be 5409 contained in its regulations, the contract, or the grant), the 5410 author retains copyright. If a provision of the contract, grant, or 5411 regulation states that there will be no copyright, then it does not 5412 exist. When a work is created, copyright automatically comes into 5413 existence unless something exists that says it does not. 5414 5415 * An enhanced electronic copy of a print copy of an older reference 5416 work in the public domain that does not contain copyrightable new 5417 material is a purely mechanical rendition of the original work, and 5418 is not copyrightable. 5419 5420 * Usually, when a work enters the public domain, nothing can remove 5421 it. For example, Congress recently passed into law the concept of 5422 automatic renewal, which means that copyright on any work published 5423 between l964 and l978 does not have to be renewed in order to 5424 receive a seventy-five-year term. But any work not renewed before 5425 1964 is in the public domain. 5426 5427 * Concerning whether or not the United States keeps track of when 5428 authors die, nothing was ever done, nor is anything being done at 5429 the moment by the Copyright Office. 5430 5431 * Software that drives a mechanical process is itself copyrightable. 5432 If one changes platforms, the software itself has a copyright. The 5433 World Intellectual Property Organization will hold a symposium 28 5434 March through 2 April l993, at Harvard University, on digital 5435 technology, and will study this entire issue. If one purchases a 5436 computer software package, such as MacPaint, and creates something 5437 new, one receives protection only for that which has been added. 5438 5439 PETERS added that often in copyright matters, rough justice is the 5440 outcome, for example, in collective licensing, ASCAP (i.e., American 5441 Society of Composers, Authors, and Publishers), and BMI (i.e., Broadcast 5442 Music, Inc.), where it may seem that the big guys receive more than their 5443 due. Of course, people ought not to copy a creative product without 5444 paying for it; there should be some compensation. But the truth of the 5445 world, and it is not a great truth, is that the big guy gets played on 5446 the radio more frequently than the little guy, who has to do much more 5447 until he becomes a big guy. That is true of every author, every 5448 composer, everyone, and, unfortunately, is part of life. 5449 5450 Copyright always originates with the author, except in cases of works 5451 made for hire. (Most software falls into this category.) When an author 5452 sends his article to a journal, he has not relinquished copyright, though 5453 he retains the right to relinquish it. The author receives absolutely 5454 everything. The less prominent the author, the more leverage the 5455 publisher will have in contract negotiations. In order to transfer the 5456 rights, the author must sign an agreement giving them away. 5457 5458 In an electronic society, it is important to be able to license a writer 5459 and work out deals. With regard to use of a work, it usually is much 5460 easier when a publisher holds the rights. In an electronic era, a real 5461 problem arises when one is digitizing and making information available. 5462 PETERS referred again to electronic licensing clearinghouses. Copyright 5463 ought to remain with the author, but as one moves forward globally in the 5464 electronic arena, a middleman who can handle the various rights becomes 5465 increasingly necessary. 5466 5467 The notion of copyright law is that it resides with the individual, but 5468 in an on-line environment, where a work can be adapted and tinkered with 5469 by many individuals, there is concern. If changes are authorized and 5470 there is no agreement to the contrary, the person who changes a work owns 5471 the changes. To put it another way, the person who acquires permission 5472 to change a work technically will become the author and the owner, unless 5473 some agreement to the contrary has been made. It is typical for the 5474 original publisher to try to control all of the versions and all of the 5475 uses. Copyright law always only sets up the boundaries. Anything can be 5476 changed by contract. 5477 5478 ****** 5479 5480 SESSION VII. CONCLUSION 5481 5482 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 5483 GENERAL DISCUSSION * Two questions for discussion * Different emphases in 5484 the Workshop * Bringing the text and image partisans together * 5485 Desiderata in planning the long-term development of something * Questions 5486 surrounding the issue of electronic deposit * Discussion of electronic 5487 deposit as an allusion to the issue of standards * Need for a directory 5488 of preservation projects in digital form and for access to their 5489 digitized files * CETH's catalogue of machine-readable texts in the 5490 humanities * What constitutes a publication in the electronic world? * 5491 Need for LC to deal with the concept of on-line publishing * LC's Network 5492 Development Office exploring the limits of MARC as a standard in terms 5493 of handling electronic information * Magnitude of the problem and the 5494 need for distributed responsibility in order to maintain and store 5495 electronic information * Workshop participants to be viewed as a starting 5496 point * Development of a network version of AM urged * A step toward AM's 5497 construction of some sort of apparatus for network access * A delicate 5498 and agonizing policy question for LC * Re the issue of electronic 5499 deposit, LC urged to initiate a catalytic process in terms of distributed 5500 responsibility * Suggestions for cooperative ventures * Commercial 5501 publishers' fears * Strategic questions for getting the image and text 5502 people to think through long-term cooperation * Clarification of the 5503 driving force behind both the Perseus and the Cornell Xerox projects * 5504 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 5505 5506 In his role as moderator of the concluding session, GIFFORD raised two 5507 questions he believed would benefit from discussion: 1) Are there enough 5508 commonalities among those of us that have been here for two days so that 5509 we can see courses of action that should be taken in the future? And, if 5510 so, what are they and who might take them? 2) Partly derivative from 5511 that, but obviously very dangerous to LC as host, do you see a role for 5512 the Library of Congress in all this? Of course, the Library of Congress 5513 holds a rather special status in a number of these matters, because it is 5514 not perceived as a player with an economic stake in them, but are there 5515 roles that LC can play that can help advance us toward where we are heading? 5516 5517 Describing himself as an uninformed observer of the technicalities of the 5518 last two days, GIFFORD detected three different emphases in the Workshop: 5519 1) people who are very deeply committed to text; 2) people who are almost 5520 passionate about images; and 3) a few people who are very committed to 5521 what happens to the networks. In other words, the new networking 5522 dimension, the accessibility of the processability, the portability of 5523 all this across the networks. How do we pull those three together? 5524 5525 Adding a question that reflected HOCKEY's comment that this was the 5526 fourth workshop she had attended in the previous thirty days, FLEISCHHAUER 5527 wondered to what extent this meeting had reinvented the wheel, or if it 5528 had contributed anything in the way of bringing together a different group 5529 of people from those who normally appear on the workshop circuit. 5530 5531 HOCKEY confessed to being struck at this meeting and the one the 5532 Electronic Pierce Consortium organized the previous week that this was a 5533 coming together of people working on texts and not images. Attempting to 5534 bring the two together is something we ought to be thinking about for the 5535 future: How one can think about working with image material to begin 5536 with, but structuring it and digitizing it in such a way that at a later 5537 stage it can be interpreted into text, and find a common way of building 5538 text and images together so that they can be used jointly in the future, 5539 with the network support to begin there because that is how people will 5540 want to access it. 5541 5542 In planning the long-term development of something, which is what is 5543 being done in electronic text, HOCKEY stressed the importance not only 5544 of discussing the technical aspects of how one does it but particularly 5545 of thinking about what the people who use the stuff will want to do. 5546 But conversely, there are numerous things that people start to do with 5547 electronic text or material that nobody ever thought of in the beginning. 5548 5549 LESK, in response to the question concerning the role of the Library of 5550 Congress, remarked the often suggested desideratum of having electronic 5551 deposit: Since everything is now computer-typeset, an entire decade of 5552 material that was machine-readable exists, but the publishers frequently 5553 did not save it; has LC taken any action to have its copyright deposit 5554 operation start collecting these machine-readable versions? In the 5555 absence of PETERS, GIFFORD replied that the question was being 5556 actively considered but that that was only one dimension of the problem. 5557 Another dimension is the whole question of the integrity of the original 5558 electronic document. It becomes highly important in science to prove 5559 authorship. How will that be done? 5560 5561 ERWAY explained that, under the old policy, to make a claim for a 5562 copyright for works that were published in electronic form, including 5563 software, one had to submit a paper copy of the first and last twenty 5564 pages of code--something that represented the work but did not include 5565 the entire work itself and had little value to anyone. As a temporary 5566 measure, LC has claimed the right to demand electronic versions of 5567 electronic publications. This measure entails a proactive role for the 5568 Library to say that it wants a particular electronic version. Publishers 5569 then have perhaps a year to submit it. But the real problem for LC is 5570 what to do with all this material in all these different formats. Will 5571 the Library mount it? How will it give people access to it? How does LC 5572 keep track of the appropriate computers, software, and media? The situation 5573 is so hard to control, ERWAY said, that it makes sense for each publishing 5574 house to maintain its own archive. But LC cannot enforce that either. 5575 5576 GIFFORD acknowledged LESK's suggestion that establishing a priority 5577 offered the solution, albeit a fairly complicated one. But who maintains 5578 that register?, he asked. GRABER noted that LC does attempt to collect a 5579 Macintosh version and the IBM-compatible version of software. It does 5580 not collect other versions. But while true for software, BYRUM observed, 5581 this reply does not speak to materials, that is, all the materials that 5582 were published that were on somebody's microcomputer or driver tapes 5583 at a publishing office across the country. LC does well to acquire 5584 specific machine-readable products selectively that were intended to be 5585 machine-readable. Materials that were in machine-readable form at one time, 5586 BYRUM said, would be beyond LC's capability at the moment, insofar as 5587 attempting to acquire, organize, and preserve them are concerned--and 5588 preservation would be the most important consideration. In this 5589 connection, GIFFORD reiterated the need to work out some sense of 5590 distributive responsibility for a number of these issues, which 5591 inevitably will require significant cooperation and discussion. 5592 Nobody can do it all. 5593 5594 LESK suggested that some publishers may look with favor on LC beginning 5595 to serve as a depository of tapes in an electronic manuscript standard. 5596 Publishers may view this as a service that they did not have to perform 5597 and they might send in tapes. However, SPERBERG-McQUEEN countered, 5598 although publishers have had equivalent services available to them for a 5599 long time, the electronic text archive has never turned away or been 5600 flooded with tapes and is forever sending feedback to the depositor. 5601 Some publishers do send in tapes. 5602 5603 ANDRE viewed this discussion as an allusion to the issue of standards. 5604 She recommended that the AAP standard and the TEI, which has already been 5605 somewhat harmonized internationally and which also shares several 5606 compatibilities with the AAP, be harmonized to ensure sufficient 5607 compatibility in the software. She drew the line at saying LC ought to 5608 be the locus or forum for such harmonization. 5609 5610 Taking the group in a slightly different direction, but one where at 5611 least in the near term LC might play a helpful role, LYNCH remarked the 5612 plans of a number of projects to carry out preservation by creating 5613 digital images that will end up in on-line or near-line storage at some 5614 institution. Presumably, LC will link this material somehow to its 5615 on-line catalog in most cases. Thus, it is in a digital form. LYNCH had 5616 the impression that many of these institutions would be willing to make 5617 those files accessible to other people outside the institution, provided 5618 that there is no copyright problem. This desideratum will require 5619 propagating the knowledge that those digitized files exist, so that they 5620 can end up in other on-line catalogs. Although uncertain about the 5621 mechanism for achieving this result, LYNCH said that it warranted 5622 scrutiny because it seemed to be connected to some of the basic issues of 5623 cataloging and distribution of records. It would be foolish, given the 5624 amount of work that all of us have to do and our meager resources, to 5625 discover multiple institutions digitizing the same work. Re microforms, 5626 LYNCH said, we are in pretty good shape. 5627 5628 BATTIN called this a big problem and noted that the Cornell people (who 5629 had already departed) were working on it. At issue from the beginning 5630 was to learn how to catalog that information into RLIN and then into 5631 OCLC, so that it would be accessible. That issue remains to be resolved. 5632 LYNCH rejoined that putting it into OCLC or RLIN was helpful insofar as 5633 somebody who is thinking of performing preservation activity on that work 5634 could learn about it. It is not necessarily helpful for institutions to 5635 make that available. BATTIN opined that the idea was that it not only be 5636 for preservation purposes but for the convenience of people looking for 5637 this material. She endorsed LYNCH's dictum that duplication of this 5638 effort was to be avoided by every means. 5639 5640 HOCKEY informed the Workshop about one major current activity of CETH, 5641 namely a catalogue of machine-readable texts in the humanities. Held on 5642 RLIN at present, the catalogue has been concentrated on ASCII as opposed 5643 to digitized images of text. She is exploring ways to improve the 5644 catalogue and make it more widely available, and welcomed suggestions 5645 about these concerns. CETH owns the records, which are not just 5646 restricted to RLIN, and can distribute them however it wishes. 5647 5648 Taking up LESK's earlier question, BATTIN inquired whether LC, since it 5649 is accepting electronic files and designing a mechanism for dealing with 5650 that rather than putting books on shelves, would become responsible for 5651 the National Copyright Depository of Electronic Materials. Of course 5652 that could not be accomplished overnight, but it would be something LC 5653 could plan for. GIFFORD acknowledged that much thought was being devoted 5654 to that set of problems and returned the discussion to the issue raised 5655 by LYNCH--whether or not putting the kind of records that both BATTIN and 5656 HOCKEY have been talking about in RLIN is not a satisfactory solution. 5657 It seemed to him that RLIN answered LYNCH's original point concerning 5658 some kind of directory for these kinds of materials. In a situation 5659 where somebody is attempting to decide whether or not to scan this or 5660 film that or to learn whether or not someone has already done so, LYNCH 5661 suggested, RLIN is helpful, but it is not helpful in the case of a local, 5662 on-line catalogue. Further, one would like to have her or his system be 5663 aware that that exists in digital form, so that one can present it to a 5664 patron, even though one did not digitize it, if it is out of copyright. 5665 The only way to make those linkages would be to perform a tremendous 5666 amount of real-time look-up, which would be awkward at best, or 5667 periodically to yank the whole file from RLIN and match it against one's 5668 own stuff, which is a nuisance. 5669 5670 But where, ERWAY inquired, does one stop including things that are 5671 available with Internet, for instance, in one's local catalogue? 5672 It almost seems that that is LC's means to acquire access to them. 5673 That represents LC's new form of library loan. Perhaps LC's new on-line 5674 catalogue is an amalgamation of all these catalogues on line. LYNCH 5675 conceded that perhaps that was true in the very long term, but was not 5676 applicable to scanning in the short term. In his view, the totals cited 5677 by Yale, 10,000 books over perhaps a four-year period, and 1,000-1,500 5678 books from Cornell, were not big numbers, while searching all over 5679 creation for relatively rare occurrences will prove to be less efficient. 5680 As GIFFORD wondered if this would not be a separable file on RLIN and 5681 could be requested from them, BATTIN interjected that it was easily 5682 accessible to an institution. SEVERTSON pointed out that that file, cum 5683 enhancements, was available with reference information on CD-ROM, which 5684 makes it a little more available. 5685 5686 In HOCKEY's view, the real question facing the Workshop is what to put in 5687 this catalogue, because that raises the question of what constitutes a 5688 publication in the electronic world. (WEIBEL interjected that Eric Joule 5689 in OCLC's Office of Research is also wrestling with this particular 5690 problem, while GIFFORD thought it sounded fairly generic.) HOCKEY 5691 contended that a majority of texts in the humanities are in the hands 5692 of either a small number of large research institutions or individuals 5693 and are not generally available for anyone else to access at all. 5694 She wondered if these texts ought to be catalogued. 5695 5696 After argument proceeded back and forth for several minutes over why 5697 cataloguing might be a necessary service, LEBRON suggested that this 5698 issue involved the responsibility of a publisher. The fact that someone 5699 has created something electronically and keeps it under his or her 5700 control does not constitute publication. Publication implies 5701 dissemination. While it would be important for a scholar to let other 5702 people know that this creation exists, in many respects this is no 5703 different from an unpublished manuscript. That is what is being accessed 5704 in there, except that now one is not looking at it in the hard-copy but 5705 in the electronic environment. 5706 5707 LEBRON expressed puzzlement at the variety of ways electronic publishing 5708 has been viewed. Much of what has been discussed throughout these two 5709 days has concerned CD-ROM publishing, whereas in the on-line environment 5710 that she confronts, the constraints and challenges are very different. 5711 Sooner or later LC will have to deal with the concept of on-line 5712 publishing. Taking up the comment ERWAY made earlier about storing 5713 copies, LEBRON gave her own journal as an example. How would she deposit 5714 OJCCT for copyright?, she asked, because the journal will exist in the 5715 mainframe at OCLC and people will be able to access it. Here the 5716 situation is different, ownership versus access, and is something that 5717 arises with publication in the on-line environment, faster than is 5718 sometimes realized. Lacking clear answers to all of these questions 5719 herself, LEBRON did not anticipate that LC would be able to take a role 5720 in helping to define some of them for quite a while. 5721 5722 GREENFIELD observed that LC's Network Development Office is attempting, 5723 among other things, to explore the limits of MARC as a standard in terms 5724 of handling electronic information. GREENFIELD also noted that Rebecca 5725 GUENTHER from that office gave a paper to the American Society for 5726 Information Science (ASIS) summarizing several of the discussion papers 5727 that were coming out of the Network Development Office. GREENFIELD said 5728 he understood that that office had a list-server soliciting just the kind 5729 of feedback received today concerning the difficulties of identifying and 5730 cataloguing electronic information. GREENFIELD hoped that everybody 5731 would be aware of that and somehow contribute to that conversation. 5732 5733 Noting two of LC's roles, first, to act as a repository of record for 5734 material that is copyrighted in this country, and second, to make 5735 materials it holds available in some limited form to a clientele that 5736 goes beyond Congress, BESSER suggested that it was incumbent on LC to 5737 extend those responsibilities to all the things being published in 5738 electronic form. This would mean eventually accepting electronic 5739 formats. LC could require that at some point they be in a certain 5740 limited set of formats, and then develop mechanisms for allowing people 5741 to access those in the same way that other things are accessed. This 5742 does not imply that they are on the network and available to everyone. 5743 LC does that with most of its bibliographic records, BESSER said, which 5744 end up migrating to the utility (e.g., OCLC) or somewhere else. But just 5745 as most of LC's books are available in some form through interlibrary 5746 loan or some other mechanism, so in the same way electronic formats ought 5747 to be available to others in some format, though with some copyright 5748 considerations. BESSER was not suggesting that these mechanisms be 5749 established tomorrow, only that they seemed to fall within LC's purview, 5750 and that there should be long-range plans to establish them. 5751 5752 Acknowledging that those from LC in the room agreed with BESSER 5753 concerning the need to confront difficult questions, GIFFORD underscored 5754 the magnitude of the problem of what to keep and what to select. GIFFORD 5755 noted that LC currently receives some 31,000 items per day, not counting 5756 electronic materials, and argued for much more distributed responsibility 5757 in order to maintain and store electronic information. 5758 5759 BESSER responded that the assembled group could be viewed as a starting 5760 point, whose initial operating premise could be helping to move in this 5761 direction and defining how LC could do so, for example, in areas of 5762 standardization or distribution of responsibility. 5763 5764 FLEISCHHAUER added that AM was fully engaged, wrestling with some of the 5765 questions that pertain to the conversion of older historical materials, 5766 which would be one thing that the Library of Congress might do. Several 5767 points mentioned by BESSER and several others on this question have a 5768 much greater impact on those who are concerned with cataloguing and the 5769 networking of bibliographic information, as well as preservation itself. 5770 5771 Speaking directly to AM, which he considered was a largely uncopyrighted 5772 database, LYNCH urged development of a network version of AM, or 5773 consideration of making the data in it available to people interested in 5774 doing network multimedia. On account of the current great shortage of 5775 digital data that is both appealing and unencumbered by complex rights 5776 problems, this course of action could have a significant effect on making 5777 network multimedia a reality. 5778 5779 In this connection, FLEISCHHAUER reported on a fragmentary prototype in 5780 LC's Office of Information Technology Services that attempts to associate 5781 digital images of photographs with cataloguing information in ways that 5782 work within a local area network--a step, so to say, toward AM's 5783 construction of some sort of apparatus for access. Further, AM has 5784 attempted to use standard data forms in order to help make that 5785 distinction between the access tools and the underlying data, and thus 5786 believes that the database is networkable. 5787 5788 A delicate and agonizing policy question for LC, however, which comes 5789 back to resources and unfortunately has an impact on this, is to find 5790 some appropriate, honorable, and legal cost-recovery possibilities. A 5791 certain skittishness concerning cost-recovery has made people unsure 5792 exactly what to do. AM would be highly receptive to discussing further 5793 LYNCH's offer to test or demonstrate its database in a network 5794 environment, FLEISCHHAUER said. 5795 5796 Returning the discussion to what she viewed as the vital issue of 5797 electronic deposit, BATTIN recommended that LC initiate a catalytic 5798 process in terms of distributed responsibility, that is, bring together 5799 the distributed organizations and set up a study group to look at all 5800 these issues and see where we as a nation should move. The broader 5801 issues of how we deal with the management of electronic information will 5802 not disappear, but only grow worse. 5803 5804 LESK took up this theme and suggested that LC attempt to persuade one 5805 major library in each state to deal with its state equivalent publisher, 5806 which might produce a cooperative project that would be equitably 5807 distributed around the country, and one in which LC would be dealing with 5808 a minimal number of publishers and minimal copyright problems. 5809 5810 GRABER remarked the recent development in the scientific community of a 5811 willingness to use SGML and either deposit or interchange on a fairly 5812 standardized format. He wondered if a similar movement was taking place 5813 in the humanities. Although the National Library of Medicine found only 5814 a few publishers to cooperate in a like venture two or three years ago, a 5815 new effort might generate a much larger number willing to cooperate. 5816 5817 KIMBALL recounted his unit's (Machine-Readable Collections Reading Room) 5818 troubles with the commercial publishers of electronic media in acquiring 5819 materials for LC's collections, in particular the publishers' fear that 5820 they would not be able to cover their costs and would lose control of 5821 their products, that LC would give them away or sell them and make 5822 profits from them. He doubted that the publishing industry was prepared 5823 to move into this area at the moment, given its resistance to allowing LC 5824 to use its machine-readable materials as the Library would like. 5825 5826 The copyright law now addresses compact disk as a medium, and LC can 5827 request one copy of that, or two copies if it is the only version, and 5828 can request copies of software, but that fails to address magazines or 5829 books or anything like that which is in machine-readable form. 5830 5831 GIFFORD acknowledged the thorny nature of this issue, which he illustrated 5832 with the example of the cumbersome process involved in putting a copy of a 5833 scientific database on a LAN in LC's science reading room. He also 5834 acknowledged that LC needs help and could enlist the energies and talents 5835 of Workshop participants in thinking through a number of these problems. 5836 5837 GIFFORD returned the discussion to getting the image and text people to 5838 think through together where they want to go in the long term. MYLONAS 5839 conceded that her experience at the Pierce Symposium the previous week at 5840 Georgetown University and this week at LC had forced her to reevaluate 5841 her perspective on the usefulness of text as images. MYLONAS framed the 5842 issues in a series of questions: How do we acquire machine-readable 5843 text? Do we take pictures of it and perform OCR on it later? Is it 5844 important to obtain very high-quality images and text, etc.? 5845 FLEISCHHAUER agreed with MYLONAS's framing of strategic questions, adding 5846 that a large institution such as LC probably has to do all of those 5847 things at different times. Thus, the trick is to exercise judgment. The 5848 Workshop had added to his and AM's considerations in making those 5849 judgments. Concerning future meetings or discussions, MYLONAS suggested 5850 that screening priorities would be helpful. 5851 5852 WEIBEL opined that the diversity reflected in this group was a sign both 5853 of the health and of the immaturity of the field, and more time would 5854 have to pass before we convince one another concerning standards. 5855 5856 An exchange between MYLONAS and BATTIN clarified the point that the 5857 driving force behind both the Perseus and the Cornell Xerox projects was 5858 the preservation of knowledge for the future, not simply for particular 5859 research use. In the case of Perseus, MYLONAS said, the assumption was 5860 that the texts would not be entered again into electronically readable 5861 form. SPERBERG-McQUEEN added that a scanned image would not serve as an 5862 archival copy for purposes of preservation in the case of, say, the Bill 5863 of Rights, in the sense that the scanned images are effectively the 5864 archival copies for the Cornell mathematics books. 5865 5866 5867 *** *** *** ****** *** *** *** 5868 5869 5870 Appendix I: PROGRAM 5871 5872 5873 5874 WORKSHOP 5875 ON 5876 ELECTRONIC 5877 TEXTS 5878 5879 5880 5881 9-10 June 1992 5882 5883 Library of Congress 5884 Washington, D.C. 5885 5886 5887 5888 Supported by a Grant from the David and Lucile Packard Foundation 5889 5890 5891 Tuesday, 9 June 1992 5892 5893 NATIONAL DEMONSTRATION LAB, ATRIUM, LIBRARY MADISON 5894 5895 8:30 AM Coffee and Danish, registration 5896 5897 9:00 AM Welcome 5898 5899 Prosser Gifford, Director for Scholarly Programs, and Carl 5900 Fleischhauer, Coordinator, American Memory, Library of 5901 Congress 5902 5903 9:l5 AM Session I. Content in a New Form: Who Will Use It and What 5904 Will They Do? 5905 5906 Broad description of the range of electronic information. 5907 Characterization of who uses it and how it is or may be used. 5908 In addition to a look at scholarly uses, this session will 5909 include a presentation on use by students (K-12 and college) 5910 and the general public. 5911 5912 Moderator: James Daly 5913 Avra Michelson, Archival Research and Evaluation Staff, 5914 National Archives and Records Administration (Overview) 5915 Susan H. Veccia, Team Leader, American Memory, User Evaluation, 5916 and 5917 Joanne Freeman, Associate Coordinator, American Memory, Library 5918 of Congress (Beyond the scholar) 5919 5920 10:30- 5921 11:00 AM Break 5922 5923 11:00 AM Session II. Show and Tell. 5924 5925 Each presentation to consist of a fifteen-minute 5926 statement/show; group discussion will follow lunch. 5927 5928 Moderator: Jacqueline Hess, Director, National Demonstration 5929 Lab 5930 5931 1. A classics project, stressing texts and text retrieval 5932 more than multimedia: Perseus Project, Harvard 5933 University 5934 Elli Mylonas, Managing Editor 5935 5936 2. Other humanities projects employing the emerging norms of 5937 the Text Encoding Initiative (TEI): Chadwyck-Healey's 5938 The English Poetry Full Text Database and/or Patrologia 5939 Latina Database 5940 Eric M. Calaluca, Vice President, Chadwyck-Healey, Inc. 5941 5942 3. American Memory 5943 Carl Fleischhauer, Coordinator, and 5944 Ricky Erway, Associate Coordinator, Library of Congress 5945 5946 4. Founding Fathers example from Packard Humanities 5947 Institute: The Papers of George Washington, University 5948 of Virginia 5949 Dorothy Twohig, Managing Editor, and/or 5950 David Woodley Packard 5951 5952 5. An electronic medical journal offering graphics and 5953 full-text searchability: The Online Journal of Current 5954 Clinical Trials, American Association for the Advancement 5955 of Science 5956 Maria L. Lebron, Managing Editor 5957 5958 6. A project that offers facsimile images of pages but omits 5959 searchable text: Cornell math books 5960 Lynne K. Personius, Assistant Director, Cornell 5961 Information Technologies for Scholarly Information 5962 Sources, Cornell University 5963 5964 12:30 PM Lunch (Dining Room A, Library Madison 620. Exhibits 5965 available.) 5966 5967 1:30 PM Session II. Show and Tell (Cont'd.). 5968 5969 3:00- 5970 3:30 PM Break 5971 5972 3:30- 5973 5:30 PM Session III. Distribution, Networks, and Networking: Options 5974 for Dissemination. 5975 5976 Published disks: University presses and public-sector 5977 publishers, private-sector publishers 5978 Computer networks 5979 5980 Moderator: Robert G. Zich, Special Assistant to the Associate 5981 Librarian for Special Projects, Library of Congress 5982 Clifford A. Lynch, Director, Library Automation, University of 5983 California 5984 Howard Besser, School of Library and Information Science, 5985 University of Pittsburgh 5986 Ronald L. Larsen, Associate Director of Libraries for 5987 Information Technology, University of Maryland at College 5988 Park 5989 Edwin B. Brownrigg, Executive Director, Memex Research 5990 Institute 5991 5992 6:30 PM Reception (Montpelier Room, Library Madison 619.) 5993 5994 ****** 5995 5996 Wednesday, 10 June 1992 5997 5998 DINING ROOM A, LIBRARY MADISON 620 5999 6000 8:30 AM Coffee and Danish 6001 6002 9:00 AM Session IV. Image Capture, Text Capture, Overview of Text and 6003 Image Storage Formats. 6004 6005 Moderator: William L. Hooton, Vice President of Operations, 6006 I-NET 6007 6008 A) Principal Methods for Image Capture of Text: 6009 Direct scanning 6010 Use of microform 6011 6012 Anne R. Kenney, Assistant Director, Department of Preservation 6013 and Conservation, Cornell University 6014 Pamela Q.J. Andre, Associate Director, Automation, and 6015 Judith A. Zidar, Coordinator, National Agricultural Text 6016 Digitizing Program (NATDP), National Agricultural Library 6017 (NAL) 6018 Donald J. Waters, Head, Systems Office, Yale University Library 6019 6020 B) Special Problems: 6021 Bound volumes 6022 Conservation 6023 Reproducing printed halftones 6024 6025 Carl Fleischhauer, Coordinator, American Memory, Library of 6026 Congress 6027 George Thoma, Chief, Communications Engineering Branch, 6028 National Library of Medicine (NLM) 6029 6030 10:30- 6031 11:00 AM Break 6032 6033 11:00 AM Session IV. Image Capture, Text Capture, Overview of Text and 6034 Image Storage Formats (Cont'd.). 6035 6036 C) Image Standards and Implications for Preservation 6037 6038 Jean Baronas, Senior Manager, Department of Standards and 6039 Technology, Association for Information and Image Management 6040 (AIIM) 6041 Patricia Battin, President, The Commission on Preservation and 6042 Access (CPA) 6043 6044 D) Text Conversion: 6045 OCR vs. rekeying 6046 Standards of accuracy and use of imperfect texts 6047 Service bureaus 6048 6049 Stuart Weibel, Senior Research Specialist, Online Computer 6050 Library Center, Inc. (OCLC) 6051 Michael Lesk, Executive Director, Computer Science Research, 6052 Bellcore 6053 Ricky Erway, Associate Coordinator, American Memory, Library of 6054 Congress 6055 Pamela Q.J. Andre, Associate Director, Automation, and 6056 Judith A. Zidar, Coordinator, National Agricultural Text 6057 Digitizing Program (NATDP), National Agricultural Library 6058 (NAL) 6059 6060 12:30- 6061 1:30 PM Lunch 6062 6063 1:30 PM Session V. Approaches to Preparing Electronic Texts. 6064 6065 Discussion of approaches to structuring text for the computer; 6066 pros and cons of text coding, description of methods in 6067 practice, and comparison of text-coding methods. 6068 6069 Moderator: Susan Hockey, Director, Center for Electronic Texts 6070 in the Humanities (CETH), Rutgers and Princeton Universities 6071 David Woodley Packard 6072 C.M. Sperberg-McQueen, Editor, Text Encoding Initiative (TEI), 6073 University of Illinois-Chicago 6074 Eric M. Calaluca, Vice President, Chadwyck-Healey, Inc. 6075 6076 3:30- 6077 4:00 PM Break 6078 6079 4:00 PM Session VI. Copyright Issues. 6080 6081 Marybeth Peters, Policy Planning Adviser to the Register of 6082 Copyrights, Library of Congress 6083 6084 5:00 PM Session VII. Conclusion. 6085 6086 General discussion. 6087 What topics were omitted or given short shrift that anyone 6088 would like to talk about now? 6089 Is there a "group" here? What should the group do next, if 6090 anything? What should the Library of Congress do next, if 6091 anything? 6092 Moderator: Prosser Gifford, Director for Scholarly Programs, 6093 Library of Congress 6094 6095 6:00 PM Adjourn 6096 6097 6098 *** *** *** ****** *** *** *** 6099 6100 6101 Appendix II: ABSTRACTS 6102 6103 6104 SESSION I 6105 6106 Avra MICHELSON Forecasting the Use of Electronic Texts by 6107 Social Sciences and Humanities Scholars 6108 6109 This presentation explores the ways in which electronic texts are likely 6110 to be used by the non-scientific scholarly community. Many of the 6111 remarks are drawn from a report the speaker coauthored with Jeff 6112 Rothenberg, a computer scientist at The RAND Corporation. 6113 6114 The speaker assesses 1) current scholarly use of information technology 6115 and 2) the key trends in information technology most relevant to the 6116 research process, in order to predict how social sciences and humanities 6117 scholars are apt to use electronic texts. In introducing the topic, 6118 current use of electronic texts is explored broadly within the context of 6119 scholarly communication. From the perspective of scholarly 6120 communication, the work of humanities and social sciences scholars 6121 involves five processes: 1) identification of sources, 2) communication 6122 with colleagues, 3) interpretation and analysis of data, 4) dissemination 6123 of research findings, and 5) curriculum development and instruction. The 6124 extent to which computation currently permeates aspects of scholarly 6125 communication represents a viable indicator of the prospects for 6126 electronic texts. 6127 6128 The discussion of current practice is balanced by an analysis of key 6129 trends in the scholarly use of information technology. These include the 6130 trends toward end-user computing and connectivity, which provide a 6131 framework for forecasting the use of electronic texts through this 6132 millennium. The presentation concludes with a summary of the ways in 6133 which the nonscientific scholarly community can be expected to use 6134 electronic texts, and the implications of that use for information 6135 providers. 6136 6137 Susan VECCIA and Joanne FREEMAN Electronic Archives for the Public: 6138 Use of American Memory in Public and 6139 School Libraries 6140 6141 This joint discussion focuses on nonscholarly applications of electronic 6142 library materials, specifically addressing use of the Library of Congress 6143 American Memory (AM) program in a small number of public and school 6144 libraries throughout the United States. AM consists of selected Library 6145 of Congress primary archival materials, stored on optical media 6146 (CD-ROM/videodisc), and presented with little or no editing. Many 6147 collections are accompanied by electronic introductions and user's guides 6148 offering background information and historical context. Collections 6149 represent a variety of formats including photographs, graphic arts, 6150 motion pictures, recorded sound, music, broadsides and manuscripts, 6151 books, and pamphlets. 6152 6153 In 1991, the Library of Congress began a nationwide evaluation of AM in 6154 different types of institutions. Test sites include public libraries, 6155 elementary and secondary school libraries, college and university 6156 libraries, state libraries, and special libraries. Susan VECCIA and 6157 Joanne FREEMAN will discuss their observations on the use of AM by the 6158 nonscholarly community, using evidence gleaned from this ongoing 6159 evaluation effort. 6160 6161 VECCIA will comment on the overall goals of the evaluation project, and 6162 the types of public and school libraries included in this study. Her 6163 comments on nonscholarly use of AM will focus on the public library as a 6164 cultural and community institution, often bridging the gap between formal 6165 and informal education. FREEMAN will discuss the use of AM in school 6166 libraries. Use by students and teachers has revealed some broad 6167 questions about the use of electronic resources, as well as definite 6168 benefits gained by the "nonscholar." Topics will include the problem of 6169 grasping content and context in an electronic environment, the stumbling 6170 blocks created by "new" technologies, and the unique skills and interests 6171 awakened through use of electronic resources. 6172 6173 SESSION II 6174 6175 Elli MYLONAS The Perseus Project: Interactive Sources and 6176 Studies in Classical Greece 6177 6178 The Perseus Project (5) has just released Perseus 1.0, the first publicly 6179 available version of its hypertextual database of multimedia materials on 6180 classical Greece. Perseus is designed to be used by a wide audience, 6181 comprised of readers at the student and scholar levels. As such, it must 6182 be able to locate information using different strategies, and it must 6183 contain enough detail to serve the different needs of its users. In 6184 addition, it must be delivered so that it is affordable to its target 6185 audience. [These problems and the solutions we chose are described in 6186 Mylonas, "An Interface to Classical Greek Civilization," JASIS 43:2, 6187 March 1992.] 6188 6189 In order to achieve its objective, the project staff decided to make a 6190 conscious separation between selecting and converting textual, database, 6191 and image data on the one hand, and putting it into a delivery system on 6192 the other. That way, it is possible to create the electronic data 6193 without thinking about the restrictions of the delivery system. We have 6194 made a great effort to choose system-independent formats for our data, 6195 and to put as much thought and work as possible into structuring it so 6196 that the translation from paper to electronic form will enhance the value 6197 of the data. [A discussion of these solutions as of two years ago is in 6198 Elli Mylonas, Gregory Crane, Kenneth Morrell, and D. Neel Smith, "The 6199 Perseus Project: Data in the Electronic Age," in Accessing Antiquity: 6200 The Computerization of Classical Databases, J. Solomon and T. Worthen 6201 (eds.), University of Arizona Press, in press.] 6202 6203 Much of the work on Perseus is focused on collecting and converting the 6204 data on which the project is based. At the same time, it is necessary to 6205 provide means of access to the information, in order to make it usable, 6206 and them to investigate how it is used. As we learn more about what 6207 students and scholars from different backgrounds do with Perseus, we can 6208 adjust our data collection, and also modify the system to accommodate 6209 them. In creating a delivery system for general use, we have tried to 6210 avoid favoring any one type of use by allowing multiple forms of access 6211 to and navigation through the system. 6212 6213 The way text is handled exemplifies some of these principles. All text 6214 in Perseus is tagged using SGML, following the guidelines of the Text 6215 Encoding Initiative (TEI). This markup is used to index the text, and 6216 process it so that it can be imported into HyperCard. No SGML markup 6217 remains in the text that reaches the user, because currently it would be 6218 too expensive to create a system that acts on SGML in real time. 6219 However, the regularity provided by SGML is essential for verifying the 6220 content of the texts, and greatly speeds all the processing performed on 6221 them. The fact that the texts exist in SGML ensures that they will be 6222 relatively easy to port to different hardware and software, and so will 6223 outlast the current delivery platform. Finally, the SGML markup 6224 incorporates existing canonical reference systems (chapter, verse, line, 6225 etc.); indexing and navigation are based on these features. This ensures 6226 that the same canonical reference will always resolve to the same point 6227 within a text, and that all versions of our texts, regardless of delivery 6228 platform (even paper printouts) will function the same way. 6229 6230 In order to provide tools for users, the text is processed by a 6231 morphological analyzer, and the results are stored in a database. 6232 Together with the index, the Greek-English Lexicon, and the index of all 6233 the English words in the definitions of the lexicon, the morphological 6234 analyses comprise a set of linguistic tools that allow users of all 6235 levels to work with the textual information, and to accomplish different 6236 tasks. For example, students who read no Greek may explore a concept as 6237 it appears in Greek texts by using the English-Greek index, and then 6238 looking up works in the texts and translations, or scholars may do 6239 detailed morphological studies of word use by using the morphological 6240 analyses of the texts. Because these tools were not designed for any one 6241 use, the same tools and the same data can be used by both students and 6242 scholars. 6243 6244 NOTES: 6245 (5) Perseus is based at Harvard University, with collaborators at 6246 several other universities. The project has been funded primarily 6247 by the Annenberg/CPB Project, as well as by Harvard University, 6248 Apple Computer, and others. It is published by Yale University 6249 Press. Perseus runs on Macintosh computers, under the HyperCard 6250 program. 6251 6252 Eric CALALUCA 6253 6254 Chadwyck-Healey embarked last year on two distinct yet related full-text 6255 humanities database projects. 6256 6257 The English Poetry Full-Text Database and the Patrologia Latina Database 6258 represent new approaches to linguistic research resources. The size and 6259 complexity of the projects present problems for electronic publishers, 6260 but surmountable ones if they remain abreast of the latest possibilities 6261 in data capture and retrieval software techniques. 6262 6263 The issues which required address prior to the commencement of the 6264 projects were legion: 6265 6266 1. Editorial selection (or exclusion) of materials in each 6267 database 6268 6269 2. Deciding whether or not to incorporate a normative encoding 6270 structure into the databases? 6271 A. If one is selected, should it be SGML? 6272 B. If SGML, then the TEI? 6273 6274 3. Deliver as CD-ROM, magnetic tape, or both? 6275 6276 4. Can one produce retrieval software advanced enough for the 6277 postdoctoral linguist, yet accessible enough for unattended 6278 general use? Should one try? 6279 6280 5. Re fair and liberal networking policies, what are the risks to 6281 an electronic publisher? 6282 6283 6. How does the emergence of national and international education 6284 networks affect the use and viability of research projects 6285 requiring high investment? Do the new European Community 6286 directives concerning database protection necessitate two 6287 distinct publishing projects, one for North America and one for 6288 overseas? 6289 6290 From new notions of "scholarly fair use" to the future of optical media, 6291 virtually every issue related to electronic publishing was aired. The 6292 result is two projects which have been constructed to provide the quality 6293 research resources with the fewest encumbrances to use by teachers and 6294 private scholars. 6295 6296 Dorothy TWOHIG 6297 6298 In spring 1988 the editors of the papers of George Washington, John 6299 Adams, Thomas Jefferson, James Madison, and Benjamin Franklin were 6300 approached by classics scholar David Packard on behalf of the Packard 6301 Humanities Foundation with a proposal to produce a CD-ROM edition of the 6302 complete papers of each of the Founding Fathers. This electronic edition 6303 will supplement the published volumes, making the documents widely 6304 available to students and researchers at reasonable cost. We estimate 6305 that our CD-ROM edition of Washington's Papers will be substantially 6306 completed within the next two years and ready for publication. Within 6307 the next ten years or so, similar CD-ROM editions of the Franklin, Adams, 6308 Jefferson, and Madison papers also will be available. At the Library of 6309 Congress's session on technology, I would like to discuss not only the 6310 experience of the Washington Papers in producing the CD-ROM edition, but 6311 the impact technology has had on these major editorial projects. 6312 Already, we are editing our volumes with an eye to the material that will 6313 be readily available in the CD-ROM edition. The completed electronic 6314 edition will provide immense possibilities for the searching of documents 6315 for information in a way never possible before. The kind of technical 6316 innovations that are currently available and on the drawing board will 6317 soon revolutionize historical research and the production of historical 6318 documents. Unfortunately, much of this new technology is not being used 6319 in the planning stages of historical projects, simply because many 6320 historians are aware only in the vaguest way of its existence. At least 6321 two major new historical editing projects are considering microfilm 6322 editions, simply because they are not aware of the possibilities of 6323 electronic alternatives and the advantages of the new technology in terms 6324 of flexibility and research potential compared to microfilm. In fact, 6325 too many of us in history and literature are still at the stage of 6326 struggling with our PCs. There are many historical editorial projects in 6327 progress presently, and an equal number of literary projects. While the 6328 two fields have somewhat different approaches to textual editing, there 6329 are ways in which electronic technology can be of service to both. 6330 6331 Since few of the editors involved in the Founding Fathers CD-ROM editions 6332 are technical experts in any sense, I hope to point out in my discussion 6333 of our experience how many of these electronic innovations can be used 6334 successfully by scholars who are novices in the world of new technology. 6335 One of the major concerns of the sponsors of the multitude of new 6336 scholarly editions is the limited audience reached by the published 6337 volumes. Most of these editions are being published in small quantities 6338 and the publishers' price for them puts them out of the reach not only of 6339 individual scholars but of most public libraries and all but the largest 6340 educational institutions. However, little attention is being given to 6341 ways in which technology can bypass conventional publication to make 6342 historical and literary documents more widely available. 6343 6344 What attracted us most to the CD-ROM edition of The Papers of George 6345 Washington was the fact that David Packard's aim was to make a complete 6346 edition of all of the 135,000 documents we have collected available in an 6347 inexpensive format that would be placed in public libraries, small 6348 colleges, and even high schools. This would provide an audience far 6349 beyond our present 1,000-copy, $45 published edition. Since the CD-ROM 6350 edition will carry none of the explanatory annotation that appears in the 6351 published volumes, we also feel that the use of the CD-ROM will lead many 6352 researchers to seek out the published volumes. 6353 6354 In addition to ignorance of new technical advances, I have found that too 6355 many editors--and historians and literary scholars--are resistant and 6356 even hostile to suggestions that electronic technology may enhance their 6357 work. I intend to discuss some of the arguments traditionalists are 6358 advancing to resist technology, ranging from distrust of the speed with 6359 which it changes (we are already wondering what is out there that is 6360 better than CD-ROM) to suspicion of the technical language used to 6361 describe electronic developments. 6362 6363 Maria LEBRON 6364 6365 The Online Journal of Current Clinical Trials, a joint venture of the 6366 American Association for the Advancement of Science (AAAS) and the Online 6367 Computer Library Center, Inc. (OCLC), is the first peer-reviewed journal 6368 to provide full text, tabular material, and line illustrations on line. 6369 This presentation will discuss the genesis and start-up period of the 6370 journal. Topics of discussion will include historical overview, 6371 day-to-day management of the editorial peer review, and manuscript 6372 tagging and publication. A demonstration of the journal and its features 6373 will accompany the presentation. 6374 6375 Lynne PERSONIUS 6376 6377 Cornell University Library, Cornell Information Technologies, and Xerox 6378 Corporation, with the support of the Commission on Preservation and 6379 Access, and Sun Microsystems, Inc., have been collaborating in a project 6380 to test a prototype system for recording brittle books as digital images 6381 and producing, on demand, high-quality archival paper replacements. The 6382 project goes beyond that, however, to investigate some of the issues 6383 surrounding scanning, storing, retrieving, and providing access to 6384 digital images in a network environment. 6385 6386 The Joint Study in Digital Preservation began in January 1990. Xerox 6387 provided the College Library Access and Storage System (CLASS) software, 6388 a prototype 600-dots-per-inch (dpi) scanner, and the hardware necessary 6389 to support network printing on the DocuTech printer housed in Cornell's 6390 Computing and Communications Center (CCC). 6391 6392 The Cornell staff using the hardware and software became an integral part 6393 of the development and testing process for enhancements to the CLASS 6394 software system. The collaborative nature of this relationship is 6395 resulting in a system that is specifically tailored to the preservation 6396 application. 6397 6398 A digital library of 1,000 volumes (or approximately 300,000 images) has 6399 been created and is stored on an optical jukebox that resides in CCC. 6400 The library includes a collection of select mathematics monographs that 6401 provides mathematics faculty with an opportunity to use the electronic 6402 library. The remaining volumes were chosen for the library to test the 6403 various capabilities of the scanning system. 6404 6405 One project objective is to provide users of the Cornell library and the 6406 library staff with the ability to request facsimiles of digitized images 6407 or to retrieve the actual electronic image for browsing. A prototype 6408 viewing workstation has been created by Xerox, with input into the design 6409 by a committee of Cornell librarians and computer professionals. This 6410 will allow us to experiment with patron access to the images that make up 6411 the digital library. The viewing station provides search, retrieval, and 6412 (ultimately) printing functions with enhancements to facilitate 6413 navigation through multiple documents. 6414 6415 Cornell currently is working to extend access to the digital library to 6416 readers using workstations from their offices. This year is devoted to 6417 the development of a network resident image conversion and delivery 6418 server, and client software that will support readers who use Apple 6419 Macintosh computers, IBM windows platforms, and Sun workstations. 6420 Equipment for this development was provided by Sun Microsystems with 6421 support from the Commission on Preservation and Access. 6422 6423 During the show-and-tell session of the Workshop on Electronic Texts, a 6424 prototype view station will be demonstrated. In addition, a display of 6425 original library books that have been digitized will be available for 6426 review with associated printed copies for comparison. The fifteen-minute 6427 overview of the project will include a slide presentation that 6428 constitutes a "tour" of the preservation digitizing process. 6429 6430 The final network-connected version of the viewing station will provide 6431 library users with another mechanism for accessing the digital library, 6432 and will also provide the capability of viewing images directly. This 6433 will not require special software, although a powerful computer with good 6434 graphics will be needed. 6435 6436 The Joint Study in Digital Preservation has generated a great deal of 6437 interest in the library community. Unfortunately, or perhaps 6438 fortunately, this project serves to raise a vast number of other issues 6439 surrounding the use of digital technology for the preservation and use of 6440 deteriorating library materials, which subsequent projects will need to 6441 examine. Much work remains. 6442 6443 SESSION III 6444 6445 Howard BESSER Networking Multimedia Databases 6446 6447 What do we have to consider in building and distributing databases of 6448 visual materials in a multi-user environment? This presentation examines 6449 a variety of concerns that need to be addressed before a multimedia 6450 database can be set up in a networked environment. 6451 6452 In the past it has not been feasible to implement databases of visual 6453 materials in shared-user environments because of technological barriers. 6454 Each of the two basic models for multi-user multimedia databases has 6455 posed its own problem. The analog multimedia storage model (represented 6456 by Project Athena's parallel analog and digital networks) has required an 6457 incredibly complex (and expensive) infrastructure. The economies of 6458 scale that make multi-user setups cheaper per user served do not operate 6459 in an environment that requires a computer workstation, videodisc player, 6460 and two display devices for each user. 6461 6462 The digital multimedia storage model has required vast amounts of storage 6463 space (as much as one gigabyte per thirty still images). In the past the 6464 cost of such a large amount of storage space made this model a 6465 prohibitive choice as well. But plunging storage costs are finally 6466 making this second alternative viable. 6467 6468 If storage no longer poses such an impediment, what do we need to 6469 consider in building digitally stored multi-user databases of visual 6470 materials? This presentation will examine the networking and 6471 telecommunication constraints that must be overcome before such databases 6472 can become commonplace and useful to a large number of people. 6473 6474 The key problem is the vast size of multimedia documents, and how this 6475 affects not only storage but telecommunications transmission time. 6476 Anything slower than T-1 speed is impractical for files of 1 megabyte or 6477 larger (which is likely to be small for a multimedia document). For 6478 instance, even on a 56 Kb line it would take three minutes to transfer a 6479 1-megabyte file. And these figures assume ideal circumstances, and do 6480 not take into consideration other users contending for network bandwidth, 6481 disk access time, or the time needed for remote display. Current common 6482 telephone transmission rates would be completely impractical; few users 6483 would be willing to wait the hour necessary to transmit a single image at 6484 2400 baud. 6485 6486 This necessitates compression, which itself raises a number of other 6487 issues. In order to decrease file sizes significantly, we must employ 6488 lossy compression algorithms. But how much quality can we afford to 6489 lose? To date there has been only one significant study done of 6490 image-quality needs for a particular user group, and this study did not 6491 look at loss resulting from compression. Only after identifying 6492 image-quality needs can we begin to address storage and network bandwidth 6493 needs. 6494 6495 Experience with X-Windows-based applications (such as Imagequery, the 6496 University of California at Berkeley image database) demonstrates the 6497 utility of a client-server topology, but also points to the limitation of 6498 current software for a distributed environment. For example, 6499 applications like Imagequery can incorporate compression, but current X 6500 implementations do not permit decompression at the end user's 6501 workstation. Such decompression at the host computer alleviates storage 6502 capacity problems while doing nothing to address problems of 6503 telecommunications bandwidth. 6504 6505 We need to examine the effects on network through-put of moving 6506 multimedia documents around on a network. We need to examine various 6507 topologies that will help us avoid bottlenecks around servers and 6508 gateways. Experience with applications such as these raise still broader 6509 questions. How closely is the multimedia document tied to the software 6510 for viewing it? Can it be accessed and viewed from other applications? 6511 Experience with the MARC format (and more recently with the Z39.50 6512 protocols) shows how useful it can be to store documents in a form in 6513 which they can be accessed by a variety of application software. 6514 6515 Finally, from an intellectual-access standpoint, we need to address the 6516 issue of providing access to these multimedia documents in 6517 interdisciplinary environments. We need to examine terminology and 6518 indexing strategies that will allow us to provide access to this material 6519 in a cross-disciplinary way. 6520 6521 Ronald LARSEN Directions in High-Performance Networking for 6522 Libraries 6523 6524 The pace at which computing technology has advanced over the past forty 6525 years shows no sign of abating. Roughly speaking, each five-year period 6526 has yielded an order-of-magnitude improvement in price and performance of 6527 computing equipment. No fundamental hurdles are likely to prevent this 6528 pace from continuing for at least the next decade. It is only in the 6529 past five years, though, that computing has become ubiquitous in 6530 libraries, affecting all staff and patrons, directly or indirectly. 6531 6532 During these same five years, communications rates on the Internet, the 6533 principal academic computing network, have grown from 56 kbps to 1.5 6534 Mbps, and the NSFNet backbone is now running 45 Mbps. Over the next five 6535 years, communication rates on the backbone are expected to exceed 1 Gbps. 6536 Growth in both the population of network users and the volume of network 6537 traffic has continued to grow geometrically, at rates approaching 15 6538 percent per month. This flood of capacity and use, likened by some to 6539 "drinking from a firehose," creates immense opportunities and challenges 6540 for libraries. Libraries must anticipate the future implications of this 6541 technology, participate in its development, and deploy it to ensure 6542 access to the world's information resources. 6543 6544 The infrastructure for the information age is being put in place. 6545 Libraries face strategic decisions about their role in the development, 6546 deployment, and use of this infrastructure. The emerging infrastructure 6547 is much more than computers and communication lines. It is more than the 6548 ability to compute at a remote site, send electronic mail to a peer 6549 across the country, or move a file from one library to another. The next 6550 five years will witness substantial development of the information 6551 infrastructure of the network. 6552 6553 In order to provide appropriate leadership, library professionals must 6554 have a fundamental understanding of and appreciation for computer 6555 networking, from local area networks to the National Research and 6556 Education Network (NREN). This presentation addresses these 6557 fundamentals, and how they relate to libraries today and in the near 6558 future. 6559 6560 Edwin BROWNRIGG Electronic Library Visions and Realities 6561 6562 The electronic library has been a vision desired by many--and rejected by 6563 some--since Vannevar Bush coined the term memex to describe an automated, 6564 intelligent, personal information system. Variations on this vision have 6565 included Ted Nelson's Xanadau, Alan Kay's Dynabook, and Lancaster's 6566 "paperless library," with the most recent incarnation being the 6567 "Knowledge Navigator" described by John Scully of Apple. But the reality 6568 of library service has been less visionary and the leap to the electronic 6569 library has eluded universities, publishers, and information technology 6570 files. 6571 6572 The Memex Research Institute (MemRI), an independent, nonprofit research 6573 and development organization, has created an Electronic Library Program 6574 of shared research and development in order to make the collective vision 6575 more concrete. The program is working toward the creation of large, 6576 indexed publicly available electronic image collections of published 6577 documents in academic, special, and public libraries. This strategic 6578 plan is the result of the first stage of the program, which has been an 6579 investigation of the information technologies available to support such 6580 an effort, the economic parameters of electronic service compared to 6581 traditional library operations, and the business and political factors 6582 affecting the shift from print distribution to electronic networked 6583 access. 6584 6585 The strategic plan envisions a combination of publicly searchable access 6586 databases, image (and text) document collections stored on network "file 6587 servers," local and remote network access, and an intellectual property 6588 management-control system. This combination of technology and 6589 information content is defined in this plan as an E-library or E-library 6590 collection. Some participating sponsors are already developing projects 6591 based on MemRI's recommended directions. 6592 6593 The E-library strategy projected in this plan is a visionary one that can 6594 enable major changes and improvements in academic, public, and special 6595 library service. This vision is, though, one that can be realized with 6596 today's technology. At the same time, it will challenge the political 6597 and social structure within which libraries operate: in academic 6598 libraries, the traditional emphasis on local collections, extending to 6599 accreditation issues; in public libraries, the potential of electronic 6600 branch and central libraries fully available to the public; and for 6601 special libraries, new opportunities for shared collections and networks. 6602 6603 The environment in which this strategic plan has been developed is, at 6604 the moment, dominated by a sense of library limits. The continued 6605 expansion and rapid growth of local academic library collections is now 6606 clearly at an end. Corporate libraries, and even law libraries, are 6607 faced with operating within a difficult economic climate, as well as with 6608 very active competition from commercial information sources. For 6609 example, public libraries may be seen as a desirable but not critical 6610 municipal service in a time when the budgets of safety and health 6611 agencies are being cut back. 6612 6613 Further, libraries in general have a very high labor-to-cost ratio in 6614 their budgets, and labor costs are still increasing, notwithstanding 6615 automation investments. It is difficult for libraries to obtain capital, 6616 startup, or seed funding for innovative activities, and those 6617 technology-intensive initiatives that offer the potential of decreased 6618 labor costs can provoke the opposition of library staff. 6619 6620 However, libraries have achieved some considerable successes in the past 6621 two decades by improving both their service and their credibility within 6622 their organizations--and these positive changes have been accomplished 6623 mostly with judicious use of information technologies. The advances in 6624 computing and information technology have been well-chronicled: the 6625 continuing precipitous drop in computing costs, the growth of the 6626 Internet and private networks, and the explosive increase in publicly 6627 available information databases. 6628 6629 For example, OCLC has become one of the largest computer network 6630 organizations in the world by creating a cooperative cataloging network 6631 of more than 6,000 libraries worldwide. On-line public access catalogs 6632 now serve millions of users on more than 50,000 dedicated terminals in 6633 the United States alone. The University of California MELVYL on-line 6634 catalog system has now expanded into an index database reference service 6635 and supports more than six million searches a year. And, libraries have 6636 become the largest group of customers of CD-ROM publishing technology; 6637 more than 30,000 optical media publications such as those offered by 6638 InfoTrac and Silver Platter are subscribed to by U.S. libraries. 6639 6640 This march of technology continues and in the next decade will result in 6641 further innovations that are extremely difficult to predict. What is 6642 clear is that libraries can now go beyond automation of their order files 6643 and catalogs to automation of their collections themselves--and it is 6644 possible to circumvent the fiscal limitations that appear to obtain 6645 today. 6646 6647 This Electronic Library Strategic Plan recommends a paradigm shift in 6648 library service, and demonstrates the steps necessary to provide improved 6649 library services with limited capacities and operating investments. 6650 6651 SESSION IV-A 6652 6653 Anne KENNEY 6654 6655 The Cornell/Xerox Joint Study in Digital Preservation resulted in the 6656 recording of 1,000 brittle books as 600-dpi digital images and the 6657 production, on demand, of high-quality and archivally sound paper 6658 replacements. The project, which was supported by the Commission on 6659 Preservation and Access, also investigated some of the issues surrounding 6660 scanning, storing, retrieving, and providing access to digital images in 6661 a network environment. 6662 6663 Anne Kenney will focus on some of the issues surrounding direct scanning 6664 as identified in the Cornell Xerox Project. Among those to be discussed 6665 are: image versus text capture; indexing and access; image-capture 6666 capabilities; a comparison to photocopy and microfilm; production and 6667 cost analysis; storage formats, protocols, and standards; and the use of 6668 this scanning technology for preservation purposes. 6669 6670 The 600-dpi digital images produced in the Cornell Xerox Project proved 6671 highly acceptable for creating paper replacements of deteriorating 6672 originals. The 1,000 scanned volumes provided an array of image-capture 6673 challenges that are common to nineteenth-century printing techniques and 6674 embrittled material, and that defy the use of text-conversion processes. 6675 These challenges include diminished contrast between text and background, 6676 fragile and deteriorated pages, uneven printing, elaborate type faces, 6677 faint and bold text adjacency, handwritten text and annotations, nonRoman 6678 languages, and a proliferation of illustrated material embedded in text. 6679 The latter category included high-frequency and low-frequency halftones, 6680 continuous tone photographs, intricate mathematical drawings, maps, 6681 etchings, reverse-polarity drawings, and engravings. 6682 6683 The Xerox prototype scanning system provided a number of important 6684 features for capturing this diverse material. Technicians used multiple 6685 threshold settings, filters, line art and halftone definitions, 6686 autosegmentation, windowing, and software-editing programs to optimize 6687 image capture. At the same time, this project focused on production. 6688 The goal was to make scanning as affordable and acceptable as 6689 photocopying and microfilming for preservation reformatting. A 6690 time-and-cost study conducted during the last three months of this 6691 project confirmed the economic viability of digital scanning, and these 6692 findings will be discussed here. 6693 6694 From the outset, the Cornell Xerox Project was predicated on the use of 6695 nonproprietary standards and the use of common protocols when standards 6696 did not exist. Digital files were created as TIFF images which were 6697 compressed prior to storage using Group 4 CCITT compression. The Xerox 6698 software is MS DOS based and utilizes off-the shelf programs such as 6699 Microsoft Windows and Wang Image Wizard. The digital library is designed 6700 to be hardware-independent and to provide interchangeability with other 6701 institutions through network connections. Access to the digital files 6702 themselves is two-tiered: Bibliographic records for the computer files 6703 are created in RLIN and Cornell's local system and access into the actual 6704 digital images comprising a book is provided through a document control 6705 structure and a networked image file-server, both of which will be 6706 described. 6707 6708 The presentation will conclude with a discussion of some of the issues 6709 surrounding the use of this technology as a preservation tool (storage, 6710 refreshing, backup). 6711 6712 Pamela ANDRE and Judith ZIDAR 6713 6714 The National Agricultural Library (NAL) has had extensive experience with 6715 raster scanning of printed materials. Since 1987, the Library has 6716 participated in the National Agricultural Text Digitizing Project (NATDP) 6717 a cooperative effort between NAL and forty-five land grant university 6718 libraries. An overview of the project will be presented, giving its 6719 history and NAL's strategy for the future. 6720 6721 An in-depth discussion of NATDP will follow, including a description of 6722 the scanning process, from the gathering of the printed materials to the 6723 archiving of the electronic pages. The type of equipment required for a 6724 stand-alone scanning workstation and the importance of file management 6725 software will be discussed. Issues concerning the images themselves will 6726 be addressed briefly, such as image format; black and white versus color; 6727 gray scale versus dithering; and resolution. 6728 6729 Also described will be a study currently in progress by NAL to evaluate 6730 the usefulness of converting microfilm to electronic images in order to 6731 improve access. With the cooperation of Tuskegee University, NAL has 6732 selected three reels of microfilm from a collection of sixty-seven reels 6733 containing the papers, letters, and drawings of George Washington Carver. 6734 The three reels were converted into 3,500 electronic images using a 6735 specialized microfilm scanner. The selection, filming, and indexing of 6736 this material will be discussed. 6737 6738 Donald WATERS 6739 6740 Project Open Book, the Yale University Library's effort to convert 10, 6741 000 books from microfilm to digital imagery, is currently in an advanced 6742 state of planning and organization. The Yale Library has selected a 6743 major vendor to serve as a partner in the project and as systems 6744 integrator. In its proposal, the successful vendor helped isolate areas 6745 of risk and uncertainty as well as key issues to be addressed during the 6746 life of the project. The Yale Library is now poised to decide what 6747 material it will convert to digital image form and to seek funding, 6748 initially for the first phase and then for the entire project. 6749 6750 The proposal that Yale accepted for the implementation of Project Open 6751 Book will provide at the end of three phases a conversion subsystem, 6752 browsing stations distributed on the campus network within the Yale 6753 Library, a subsystem for storing 10,000 books at 200 and 600 dots per 6754 inch, and network access to the image printers. Pricing for the system 6755 implementation assumes the existence of Yale's campus ethernet network 6756 and its high-speed image printers, and includes other requisite hardware 6757 and software, as well as system integration services. Proposed operating 6758 costs include hardware and software maintenance, but do not include 6759 estimates for the facilities management of the storage devices and image 6760 servers. 6761 6762 Yale selected its vendor partner in a formal process, partly funded by 6763 the Commission for Preservation and Access. Following a request for 6764 proposal, the Yale Library selected two vendors as finalists to work with 6765 Yale staff to generate a detailed analysis of requirements for Project 6766 Open Book. Each vendor used the results of the requirements analysis to 6767 generate and submit a formal proposal for the entire project. This 6768 competitive process not only enabled the Yale Library to select its 6769 primary vendor partner but also revealed much about the state of the 6770 imaging industry, about the varying, corporate commitments to the markets 6771 for imaging technology, and about the varying organizational dynamics 6772 through which major companies are responding to and seeking to develop 6773 these markets. 6774 6775 Project Open Book is focused specifically on the conversion of images 6776 from microfilm to digital form. The technology for scanning microfilm is 6777 readily available but is changing rapidly. In its project requirements, 6778 the Yale Library emphasized features of the technology that affect the 6779 technical quality of digital image production and the costs of creating 6780 and storing the image library: What levels of digital resolution can be 6781 achieved by scanning microfilm? How does variation in the quality of 6782 microfilm, particularly in film produced to preservation standards, 6783 affect the quality of the digital images? What technologies can an 6784 operator effectively and economically apply when scanning film to 6785 separate two-up images and to control for and correct image 6786 imperfections? How can quality control best be integrated into 6787 digitizing work flow that includes document indexing and storage? 6788 6789 The actual and expected uses of digital images--storage, browsing, 6790 printing, and OCR--help determine the standards for measuring their 6791 quality. Browsing is especially important, but the facilities available 6792 for readers to browse image documents is perhaps the weakest aspect of 6793 imaging technology and most in need of development. As it defined its 6794 requirements, the Yale Library concentrated on some fundamental aspects 6795 of usability for image documents: Does the system have sufficient 6796 flexibility to handle the full range of document types, including 6797 monographs, multi-part and multivolume sets, and serials, as well as 6798 manuscript collections? What conventions are necessary to identify a 6799 document uniquely for storage and retrieval? Where is the database of 6800 record for storing bibliographic information about the image document? 6801 How are basic internal structures of documents, such as pagination, made 6802 accessible to the reader? How are the image documents physically 6803 presented on the screen to the reader? 6804 6805 The Yale Library designed Project Open Book on the assumption that 6806 microfilm is more than adequate as a medium for preserving the content of 6807 deteriorated library materials. As planning in the project has advanced, 6808 it is increasingly clear that the challenge of digital image technology 6809 and the key to the success of efforts like Project Open Book is to 6810 provide a means of both preserving and improving access to those 6811 deteriorated materials. 6812 6813 SESSION IV-B 6814 6815 George THOMA 6816 6817 In the use of electronic imaging for document preservation, there are 6818 several issues to consider, such as: ensuring adequate image quality, 6819 maintaining substantial conversion rates (through-put), providing unique 6820 identification for automated access and retrieval, and accommodating 6821 bound volumes and fragile material. 6822 6823 To maintain high image quality, image processing functions are required 6824 to correct the deficiencies in the scanned image. Some commercially 6825 available systems include these functions, while some do not. The 6826 scanned raw image must be processed to correct contrast deficiencies-- 6827 both poor overall contrast resulting from light print and/or dark 6828 background, and variable contrast resulting from stains and 6829 bleed-through. Furthermore, the scan density must be adequate to allow 6830 legibility of print and sufficient fidelity in the pseudo-halftoned gray 6831 material. Borders or page-edge effects must be removed for both 6832 compactibility and aesthetics. Page skew must be corrected for aesthetic 6833 reasons and to enable accurate character recognition if desired. 6834 Compound images consisting of both two-toned text and gray-scale 6835 illustrations must be processed appropriately to retain the quality of 6836 each. 6837 6838 SESSION IV-C 6839 6840 Jean BARONAS 6841 6842 Standards publications being developed by scientists, engineers, and 6843 business managers in Association for Information and Image Management 6844 (AIIM) standards committees can be applied to electronic image management 6845 (EIM) processes including: document (image) transfer, retrieval and 6846 evaluation; optical disk and document scanning; and document design and 6847 conversion. When combined with EIM system planning and operations, 6848 standards can assist in generating image databases that are 6849 interchangeable among a variety of systems. The applications of 6850 different approaches for image-tagging, indexing, compression, and 6851 transfer often cause uncertainty concerning EIM system compatibility, 6852 calibration, performance, and upward compatibility, until standard 6853 implementation parameters are established. The AIIM standards that are 6854 being developed for these applications can be used to decrease the 6855 uncertainty, successfully integrate imaging processes, and promote "open 6856 systems." AIIM is an accredited American National Standards Institute 6857 (ANSI) standards developer with more than twenty committees comprised of 6858 300 volunteers representing users, vendors, and manufacturers. The 6859 standards publications that are developed in these committees have 6860 national acceptance and provide the basis for international harmonization 6861 in the development of new International Organization for Standardization 6862 (ISO) standards. 6863 6864 This presentation describes the development of AIIM's EIM standards and a 6865 new effort at AIIM, a database on standards projects in a wide framework 6866 of imaging industries including capture, recording, processing, 6867 duplication, distribution, display, evaluation, and preservation. The 6868 AIIM Imagery Database will cover imaging standards being developed by 6869 many organizations in many different countries. It will contain 6870 standards publications' dates, origins, related national and 6871 international projects, status, key words, and abstracts. The ANSI Image 6872 Technology Standards Board requested that such a database be established, 6873 as did the ISO/International Electrotechnical Commission Joint Task Force 6874 on Imagery. AIIM will take on the leadership role for the database and 6875 coordinate its development with several standards developers. 6876 6877 Patricia BATTIN 6878 6879 Characteristics of standards for digital imagery: 6880 6881 * Nature of digital technology implies continuing volatility. 6882 6883 * Precipitous standard-setting not possible and probably not 6884 desirable. 6885 6886 * Standards are a complex issue involving the medium, the 6887 hardware, the software, and the technical capacity for 6888 reproductive fidelity and clarity. 6889 6890 * The prognosis for reliable archival standards (as defined by 6891 librarians) in the foreseeable future is poor. 6892 6893 Significant potential and attractiveness of digital technology as a 6894 preservation medium and access mechanism. 6895 6896 Productive use of digital imagery for preservation requires a 6897 reconceptualizing of preservation principles in a volatile, 6898 standardless world. 6899 6900 Concept of managing continuing access in the digital environment 6901 rather than focusing on the permanence of the medium and long-term 6902 archival standards developed for the analog world. 6903 6904 Transition period: How long and what to do? 6905 6906 * Redefine "archival." 6907 6908 * Remove the burden of "archival copy" from paper artifacts. 6909 6910 * Use digital technology for storage, develop management 6911 strategies for refreshing medium, hardware and software. 6912 6913 * Create acid-free paper copies for transition period backup 6914 until we develop reliable procedures for ensuring continuing 6915 access to digital files. 6916 6917 SESSION IV-D 6918 6919 Stuart WEIBEL The Role of SGML Markup in the CORE Project (6) 6920 6921 The emergence of high-speed telecommunications networks as a basic 6922 feature of the scholarly workplace is driving the demand for electronic 6923 document delivery. Three distinct categories of electronic 6924 publishing/republishing are necessary to support access demands in this 6925 emerging environment: 6926 6927 1.) Conversion of paper or microfilm archives to electronic format 6928 2.) Conversion of electronic files to formats tailored to 6929 electronic retrieval and display 6930 3.) Primary electronic publishing (materials for which the 6931 electronic version is the primary format) 6932 6933 OCLC has experimental or product development activities in each of these 6934 areas. Among the challenges that lie ahead is the integration of these 6935 three types of information stores in coherent distributed systems. 6936 6937 The CORE (Chemistry Online Retrieval Experiment) Project is a model for 6938 the conversion of large text and graphics collections for which 6939 electronic typesetting files are available (category 2). The American 6940 Chemical Society has made available computer typography files dating from 6941 1980 for its twenty journals. This collection of some 250 journal-years 6942 is being converted to an electronic format that will be accessible 6943 through several end-user applications. 6944 6945 The use of Standard Generalized Markup Language (SGML) offers the means 6946 to capture the structural richness of the original articles in a way that 6947 will support a variety of retrieval, navigation, and display options 6948 necessary to navigate effectively in very large text databases. 6949 6950 An SGML document consists of text that is marked up with descriptive tags 6951 that specify the function of a given element within the document. As a 6952 formal language construct, an SGML document can be parsed against a 6953 document-type definition (DTD) that unambiguously defines what elements 6954 are allowed and where in the document they can (or must) occur. This 6955 formalized map of article structure allows the user interface design to 6956 be uncoupled from the underlying database system, an important step 6957 toward interoperability. Demonstration of this separability is a part of 6958 the CORE project, wherein user interface designs born of very different 6959 philosophies will access the same database. 6960 6961 NOTES: 6962 (6) The CORE project is a collaboration among Cornell University's 6963 Mann Library, Bell Communications Research (Bellcore), the American 6964 Chemical Society (ACS), the Chemical Abstracts Service (CAS), and 6965 OCLC. 6966 6967 Michael LESK The CORE Electronic Chemistry Library 6968 6969 A major on-line file of chemical journal literature complete with 6970 graphics is being developed to test the usability of fully electronic 6971 access to documents, as a joint project of Cornell University, the 6972 American Chemical Society, the Chemical Abstracts Service, OCLC, and 6973 Bellcore (with additional support from Sun Microsystems, Springer-Verlag, 6974 DigitaI Equipment Corporation, Sony Corporation of America, and Apple 6975 Computers). Our file contains the American Chemical Society's on-line 6976 journals, supplemented with the graphics from the paper publication. The 6977 indexing of the articles from Chemical Abstracts Documents is available 6978 in both image and text format, and several different interfaces can be 6979 used. Our goals are (1) to assess the effectiveness and acceptability of 6980 electronic access to primary journals as compared with paper, and (2) to 6981 identify the most desirable functions of the user interface to an 6982 electronic system of journals, including in particular a comparison of 6983 page-image display with ASCII display interfaces. Early experiments with 6984 chemistry students on a variety of tasks suggest that searching tasks are 6985 completed much faster with any electronic system than with paper, but 6986 that for reading all versions of the articles are roughly equivalent. 6987 6988 Pamela ANDRE and Judith ZIDAR 6989 6990 Text conversion is far more expensive and time-consuming than image 6991 capture alone. NAL's experience with optical character recognition (OCR) 6992 will be related and compared with the experience of having text rekeyed. 6993 What factors affect OCR accuracy? How accurate does full text have to be 6994 in order to be useful? How do different users react to imperfect text? 6995 These are questions that will be explored. For many, a service bureau 6996 may be a better solution than performing the work inhouse; this will also 6997 be discussed. 6998 6999 SESSION VI 7000 7001 Marybeth PETERS 7002 7003 Copyright law protects creative works. Protection granted by the law to 7004 authors and disseminators of works includes the right to do or authorize 7005 the following: reproduce the work, prepare derivative works, distribute 7006 the work to the public, and publicly perform or display the work. In 7007 addition, copyright owners of sound recordings and computer programs have 7008 the right to control rental of their works. These rights are not 7009 unlimited; there are a number of exceptions and limitations. 7010 7011 An electronic environment places strains on the copyright system. 7012 Copyright owners want to control uses of their work and be paid for any 7013 use; the public wants quick and easy access at little or no cost. The 7014 marketplace is working in this area. Contracts, guidelines on electronic 7015 use, and collective licensing are in use and being refined. 7016 7017 Issues concerning the ability to change works without detection are more 7018 difficult to deal with. Questions concerning the integrity of the work 7019 and the status of the changed version under the copyright law are to be 7020 addressed. These are public policy issues which require informed 7021 dialogue. 7022 7023 7024 *** *** *** ****** *** *** *** 7025 7026 7027 Appendix III: DIRECTORY OF PARTICIPANTS 7028 7029 7030 PRESENTERS: 7031 7032 Pamela Q.J. Andre 7033 Associate Director, Automation 7034 National Agricultural Library 7035 10301 Baltimore Boulevard 7036 Beltsville, MD 20705-2351 7037 Phone: (301) 504-6813 7038 Fax: (301) 504-7473 7039 E-mail: INTERNET: PANDRE (a] ASRR.ARSUSDA.GOV 7040 7041 Jean Baronas, Senior Manager 7042 Department of Standards and Technology 7043 Association for Information and Image Management (AIIM) 7044 1100 Wayne Avenue, Suite 1100 7045 Silver Spring, MD 20910 7046 Phone: (301) 587-8202 7047 Fax: (301) 587-2711 7048 7049 Patricia Battin, President 7050 The Commission on Preservation and Access 7051 1400 16th Street, N.W. 7052 Suite 740 7053 Washington, DC 20036-2217 7054 Phone: (202) 939-3400 7055 Fax: (202) 939-3407 7056 E-mail: CPA (a] GWUVM.BITNET 7057 7058 Howard Besser 7059 Centre Canadien d'Architecture 7060 (Canadian Center for Architecture) 7061 1920, rue Baile 7062 Montreal, Quebec H3H 2S6 7063 CANADA 7064 Phone: (514) 939-7001 7065 Fax: (514) 939-7020 7066 E-mail: howard (a] lis.pitt.edu 7067 7068 Edwin B. Brownrigg, Executive Director 7069 Memex Research Institute 7070 422 Bonita Avenue 7071 Roseville, CA 95678 7072 Phone: (916) 784-2298 7073 Fax: (916) 786-7559 7074 E-mail: BITNET: MEMEX (a] CALSTATE.2 7075 7076 Eric M. Calaluca, Vice President 7077 Chadwyck-Healey, Inc. 7078 1101 King Street 7079 Alexandria, VA 223l4 7080 Phone: (800) 752-05l5 7081 Fax: (703) 683-7589 7082 7083 James Daly 7084 4015 Deepwood Road 7085 Baltimore, MD 21218-1404 7086 Phone: (410) 235-0763 7087 7088 Ricky Erway, Associate Coordinator 7089 American Memory 7090 Library of Congress 7091 Phone: (202) 707-6233 7092 Fax: (202) 707-3764 7093 7094 Carl Fleischhauer, Coordinator 7095 American Memory 7096 Library of Congress 7097 Phone: (202) 707-6233 7098 Fax: (202) 707-3764 7099 7100 Joanne Freeman 7101 2000 Jefferson Park Avenue, No. 7 7102 Charlottesville, VA 22903 7103 7104 Prosser Gifford 7105 Director for Scholarly Programs 7106 Library of Congress 7107 Phone: (202) 707-1517 7108 Fax: (202) 707-9898 7109 E-mail: pgif (a] seq1.loc.gov 7110 7111 Jacqueline Hess, Director 7112 National Demonstration Laboratory 7113 for Interactive Information Technologies 7114 Library of Congress 7115 Phone: (202) 707-4157 7116 Fax: (202) 707-2829 7117 7118 Susan Hockey, Director 7119 Center for Electronic Texts in the Humanities (CETH) 7120 Alexander Library 7121 Rutgers University 7122 169 College Avenue 7123 New Brunswick, NJ 08903 7124 Phone: (908) 932-1384 7125 Fax: (908) 932-1386 7126 E-mail: hockey (a] zodiac.rutgers.edu 7127 7128 William L. Hooton, Vice President 7129 Business & Technical Development 7130 Imaging & Information Systems Group 7131 I-NET 7132 6430 Rockledge Drive, Suite 400 7133 Bethesda, MD 208l7 7134 Phone: (301) 564-6750 7135 Fax: (513) 564-6867 7136 7137 Anne R. Kenney, Associate Director 7138 Department of Preservation and Conservation 7139 701 Olin Library 7140 Cornell University 7141 Ithaca, NY 14853 7142 Phone: (607) 255-6875 7143 Fax: (607) 255-9346 7144 E-mail: LYDY (a] CORNELLA.BITNET 7145 7146 Ronald L. Larsen 7147 Associate Director for Information Technology 7148 University of Maryland at College Park 7149 Room B0224, McKeldin Library 7150 College Park, MD 20742-7011 7151 Phone: (301) 405-9194 7152 Fax: (301) 314-9865 7153 E-mail: rlarsen (a] libr.umd.edu 7154 7155 Maria L. Lebron, Managing Editor 7156 The Online Journal of Current Clinical Trials 7157 l333 H Street, N.W. 7158 Washington, DC 20005 7159 Phone: (202) 326-6735 7160 Fax: (202) 842-2868 7161 E-mail: PUBSAAAS (a] GWUVM.BITNET 7162 7163 Michael Lesk, Executive Director 7164 Computer Science Research 7165 Bell Communications Research, Inc. 7166 Rm 2A-385 7167 445 South Street 7168 Morristown, NJ 07960-l9l0 7169 Phone: (201) 829-4070 7170 Fax: (201) 829-5981 7171 E-mail: lesk (a] bellcore.com (Internet) or bellcore!lesk (uucp) 7172 7173 Clifford A. Lynch 7174 Director, Library Automation 7175 University of California, 7176 Office of the President 7177 300 Lakeside Drive, 8th Floor 7178 Oakland, CA 94612-3350 7179 Phone: (510) 987-0522 7180 Fax: (510) 839-3573 7181 E-mail: calur@uccmvsa 7182 7183 Avra Michelson 7184 National Archives and Records Administration 7185 NSZ Rm. 14N 7186 7th & Pennsylvania, N.W. 7187 Washington, D.C. 20408 7188 Phone: (202) 501-5544 7189 Fax: (202) 501-5533 7190 E-mail: tmi (a] cu.nih.gov 7191 7192 Elli Mylonas, Managing Editor 7193 Perseus Project 7194 Department of the Classics 7195 Harvard University 7196 319 Boylston Hall 7197 Cambridge, MA 02138 7198 Phone: (617) 495-9025, (617) 495-0456 (direct) 7199 Fax: (617) 496-8886 7200 E-mail: Elli (a] IKAROS.Harvard.EDU or elli (a] wjh12.harvard.edu 7201 7202 David Woodley Packard 7203 Packard Humanities Institute 7204 300 Second Street, Suite 201 7205 Los Altos, CA 94002 7206 Phone: (415) 948-0150 (PHI) 7207 Fax: (415) 948-5793 7208 7209 Lynne K. Personius, Assistant Director 7210 Cornell Information Technologies for 7211 Scholarly Information Sources 7212 502 Olin Library 7213 Cornell University 7214 Ithaca, NY 14853 7215 Phone: (607) 255-3393 7216 Fax: (607) 255-9346 7217 E-mail: JRN (a] CORNELLC.BITNET 7218 7219 Marybeth Peters 7220 Policy Planning Adviser to the 7221 Register of Copyrights 7222 Library of Congress 7223 Office LM 403 7224 Phone: (202) 707-8350 7225 Fax: (202) 707-8366 7226 7227 C. Michael Sperberg-McQueen 7228 Editor, Text Encoding Initiative 7229 Computer Center (M/C 135) 7230 University of Illinois at Chicago 7231 Box 6998 7232 Chicago, IL 60680 7233 Phone: (312) 413-0317 7234 Fax: (312) 996-6834 7235 E-mail: u35395 (a] uicvm..cc.uic.edu or u35395 (a] uicvm.bitnet 7236 7237 George R. Thoma, Chief 7238 Communications Engineering Branch 7239 National Library of Medicine 7240 8600 Rockville Pike 7241 Bethesda, MD 20894 7242 Phone: (301) 496-4496 7243 Fax: (301) 402-0341 7244 E-mail: thoma (a] lhc.nlm.nih.gov 7245 7246 Dorothy Twohig, Editor 7247 The Papers of George Washington 7248 504 Alderman Library 7249 University of Virginia 7250 Charlottesville, VA 22903-2498 7251 Phone: (804) 924-0523 7252 Fax: (804) 924-4337 7253 7254 Susan H. Veccia, Team leader 7255 American Memory, User Evaluation 7256 Library of Congress 7257 American Memory Evaluation Project 7258 Phone: (202) 707-9104 7259 Fax: (202) 707-3764 7260 E-mail: svec (a] seq1.loc.gov 7261 7262 Donald J. Waters, Head 7263 Systems Office 7264 Yale University Library 7265 New Haven, CT 06520 7266 Phone: (203) 432-4889 7267 Fax: (203) 432-7231 7268 E-mail: DWATERS (a] YALEVM.BITNET or DWATERS (a] YALEVM.YCC.YALE.EDU 7269 7270 Stuart Weibel, Senior Research Scientist 7271 OCLC 7272 6565 Frantz Road 7273 Dublin, OH 43017 7274 Phone: (614) 764-608l 7275 Fax: (614) 764-2344 7276 E-mail: INTERNET: Stu (a] rsch.oclc.org 7277 7278 Robert G. Zich 7279 Special Assistant to the Associate Librarian 7280 for Special Projects 7281 Library of Congress 7282 Phone: (202) 707-6233 7283 Fax: (202) 707-3764 7284 E-mail: rzic (a] seq1.loc.gov 7285 7286 Judith A. Zidar, Coordinator 7287 National Agricultural Text Digitizing Program 7288 Information Systems Division 7289 National Agricultural Library 7290 10301 Baltimore Boulevard 7291 Beltsville, MD 20705-2351 7292 Phone: (301) 504-6813 or 504-5853 7293 Fax: (301) 504-7473 7294 E-mail: INTERNET: JZIDAR (a] ASRR.ARSUSDA.GOV 7295 7296 7297 OBSERVERS: 7298 7299 Helen Aguera, Program Officer 7300 Division of Research 7301 Room 318 7302 National Endowment for the Humanities 7303 1100 Pennsylvania Avenue, N.W. 7304 Washington, D.C. 20506 7305 Phone: (202) 786-0358 7306 Fax: (202) 786-0243 7307 7308 M. Ellyn Blanton, Deputy Director 7309 National Demonstration Laboratory 7310 for Interactive Information Technologies 7311 Library of Congress 7312 Phone: (202) 707-4157 7313 Fax: (202) 707-2829 7314 7315 Charles M. Dollar 7316 National Archives and Records Administration 7317 NSZ Rm. 14N 7318 7th & Pennsylvania, N.W. 7319 Washington, DC 20408 7320 Phone: (202) 501-5532 7321 Fax: (202) 501-5512 7322 7323 Jeffrey Field, Deputy to the Director 7324 Division of Preservation and Access 7325 Room 802 7326 National Endowment for the Humanities 7327 1100 Pennsylvania Avenue, N.W. 7328 Washington, DC 20506 7329 Phone: (202) 786-0570 7330 Fax: (202) 786-0243 7331 7332 Lorrin Garson 7333 American Chemical Society 7334 Research and Development Department 7335 1155 16th Street, N.W. 7336 Washington, D.C. 20036 7337 Phone: (202) 872-4541 7338 Fax: E-mail: INTERNET: LRG96 (a] ACS.ORG 7339 7340 William M. Holmes, Jr. 7341 National Archives and Records Administration 7342 NSZ Rm. 14N 7343 7th & Pennsylvania, N.W. 7344 Washington, DC 20408 7345 Phone: (202) 501-5540 7346 Fax: (202) 501-5512 7347 E-mail: WHOLMES (a] AMERICAN.EDU 7348 7349 Sperling Martin 7350 Information Resource Management 7351 20030 Doolittle Street 7352 Gaithersburg, MD 20879 7353 Phone: (301) 924-1803 7354 7355 Michael Neuman, Director 7356 The Center for Text and Technology 7357 Academic Computing Center 7358 238 Reiss Science Building 7359 Georgetown University 7360 Washington, DC 20057 7361 Phone: (202) 687-6096 7362 Fax: (202) 687-6003 7363 E-mail: neuman (a] guvax.bitnet, neuman (a] guvax.georgetown.edu 7364 7365 Barbara Paulson, Program Officer 7366 Division of Preservation and Access 7367 Room 802 7368 National Endowment for the Humanities 7369 1100 Pennsylvania Avenue, N.W. 7370 Washington, DC 20506 7371 Phone: (202) 786-0577 7372 Fax: (202) 786-0243 7373 7374 Allen H. Renear 7375 Senior Academic Planning Analyst 7376 Brown University Computing and Information Services 7377 115 Waterman Street 7378 Campus Box 1885 7379 Providence, R.I. 02912 7380 Phone: (401) 863-7312 7381 Fax: (401) 863-7329 7382 E-mail: BITNET: Allen@BROWNVM or 7383 INTERNET: Allen (a] brownvm.brown.edu 7384 7385 Susan M. Severtson, President 7386 Chadwyck-Healey, Inc. 7387 1101 King Street 7388 Alexandria, VA 223l4 7389 Phone: (800) 752-05l5 7390 Fax: (703) 683-7589 7391 7392 Frank Withrow 7393 U.S. Department of Education 7394 555 New Jersey Avenue, N.W. 7395 Washington, DC 20208-5644 7396 Phone: (202) 219-2200 7397 Fax: (202) 219-2106 7398 7399 7400 (LC STAFF) 7401 7402 Linda L. Arret 7403 Machine-Readable Collections Reading Room LJ 132 7404 (202) 707-1490 7405 7406 John D. Byrum, Jr. 7407 Descriptive Cataloging Division LM 540 7408 (202) 707-5194 7409 7410 Mary Jane Cavallo 7411 Science and Technology Division LA 5210 7412 (202) 707-1219 7413 7414 Susan Thea David 7415 Congressional Research Service LM 226 7416 (202) 707-7169 7417 7418 Robert Dierker 7419 Senior Adviser for Multimedia Activities LM 608 7420 (202) 707-6151 7421 7422 William W. Ellis 7423 Associate Librarian for Science and Technology LM 611 7424 (202) 707-6928 7425 7426 Ronald Gephart 7427 Manuscript Division LM 102 7428 (202) 707-5097 7429 7430 James Graber 7431 Information Technology Services LM G51 7432 (202) 707-9628 7433 7434 Rich Greenfield 7435 American Memory LM 603 7436 (202) 707-6233 7437 7438 Rebecca Guenther 7439 Network Development LM 639 7440 (202) 707-5092 7441 7442 Kenneth E. Harris 7443 Preservation LM G21 7444 (202) 707-5213 7445 7446 Staley Hitchcock 7447 Manuscript Division LM 102 7448 (202) 707-5383 7449 7450 Bohdan Kantor 7451 Office of Special Projects LM 612 7452 (202) 707-0180 7453 7454 John W. Kimball, Jr 7455 Machine-Readable Collections Reading Room LJ 132 7456 (202) 707-6560 7457 7458 Basil Manns 7459 Information Technology Services LM G51 7460 (202) 707-8345 7461 7462 Sally Hart McCallum 7463 Network Development LM 639 7464 (202) 707-6237 7465 7466 Dana J. Pratt 7467 Publishing Office LM 602 7468 (202) 707-6027 7469 7470 Jane Riefenhauser 7471 American Memory LM 603 7472 (202) 707-6233 7473 7474 William Z. Schenck 7475 Collections Development LM 650 7476 (202) 707-7706 7477 7478 Chandru J. Shahani 7479 Preservation Research and Testing Office (R&T) LM G38 7480 (202) 707-5607 7481 7482 William J. Sittig 7483 Collections Development LM 650 7484 (202) 707-7050 7485 7486 Paul Smith 7487 Manuscript Division LM 102 7488 (202) 707-5097 7489 7490 James L. Stevens 7491 Information Technology Services LM G51 7492 (202) 707-9688 7493 7494 Karen Stuart 7495 Manuscript Division LM 130 7496 (202) 707-5389 7497 7498 Tamara Swora 7499 Preservation Microfilming Office LM G05 7500 (202) 707-6293 7501 7502 Sarah Thomas 7503 Collections Cataloging LM 642 7504 (202) 707-5333 7505 7506 7507 END 7508 ************************************************************* 7509 7510 Note: This file has been edited for use on computer networks. This 7511 editing required the removal of diacritics, underlining, and fonts such 7512 as italics and bold. 7513 7514 kde 11/92 7515 7516 [A few of the italics (when used for emphasis) were replaced by CAPS mh] 7517 7518 *End of The Project Gutenberg Etext of LOC WORKSHOP ON ELECTRONIC ETEXTS 7519 7520
