Minutes of the Steering Committee Meeting Tempe, Arizona, 14-15 March 1991 C. M. Sperberg-McQueen Document Number: TEI SCM17 May 6, 1991 (18:17:02) Draft May 6, 1991 (18:17:02) Draft May 6, 1991 (18:17:02) ------------------------------------------------------------------------ Part 1 THURSDAY 14 MARCH 1 AGENDA 1. Priorities 2. Tempe meetings a. Status report session 3. Committees and Working Groups 4. Affiliated Projects 5. TEI Guidelines 6. Editors' reports on other developments 7. Usability issues 8. Software support 9. Financial issues 10. Publicity 11. Future workshops and tutorials 12. Schedule 13. Future steering committee meetings 14. Rockefeller Grant and conference 15. Establishing an enduring status for the TEI In discussing the agenda, NI urged the importance of deciding what we want to produce; SH stressed the importance of timetable and resource allocations and hoped to leave the meeting with a clear understanding of what needs to be done and how we will get it done. This subject was not proposed as a separate item on the agenda; it was implicit in the agenda passim. A propos of item 15, DW mentioned the Linguistic Data Consortium, from whose work (tending toward enduring structures) we may be able to benefit. 2 TEMPE MEETINGS Pre-conference workshop, Sunday 17 March This will be handled by LB, MSM, Elaine Brennan, and Harry Gaylord. The editors reported that a full set of overheads and a tutorial of 35 pages were ready for the workshop. The morning was to be a formal pres- entation beginning at zero, followed by very practical examples; the afternoon would be spent on demonstrations of SGML software and "toys", including a hands-on session allowing attendees to work with existing SGML software and the toys prepared by the workshop leaders. TEI Status Session, Monday 18 March MSM suggested the status session should give an overview of the project and note the current position; describe WGs and their future; describe comments and our responses. NI said it was important to focus on the future, not the past. This was agreed. SH will chair the status meeting and will decide the exact format. Perspectives session MSM, LB were asked to contact Elaine Brennan to discuss her plans for the perspectives session. We will speak if invited, and must all plan to be there, but it is at her discretion. LB volunteered to take notes during all the public meetings. 3 COMMITTEES AND WORK GROUPS Status Report on WGs TR 1 Character Sets HG has been studying Unicode. Some WSDs are under preparation. The character set WG must be very careful that Hittite, Sumerian, etc. are well taken care of. Variations in writing system within text / within words may require consultation with AI. AZ asked about Japanese, Korean, Arabic, etc. Would it be possible to have some pointers for them in the next guidelines? There is a major Japanese corpus annotated with a lot of linguistic information; it has very high economic value. AZ suggested we must do something. SH suggested including a substantial paragraph saying that examples of commonly used character sets have been given, but that others can also be used, and giving example of how to extend the normal character sets. AZ objected that character sets are included explicitly for many other languages; why not Japanese? Upon discussion it was agreed that DW should contact M. Nagao and T. Yokoi to discuss possibility of ensuring that someone from Japan be in touch with Gaylord or with other relevant WGs. DW to contact M. Nagao, T. Yokoi to discuss liaison Due: asap It was proposed that if Nagao (or anyone else accepted by DW and AZ as a reliable contact) names individuals (funded by non-TEI sources) to par- ticipate in TEI work groups, then those individuals will be welcomed as full members of the relevant work groups. Limit one per WG. Accepted. TR 2 Text Criticism Robert Kraft appears to be overburdened. Should he be replaced as chair? After discussion of several possibilities, it was agreed that MSM should approach David Chesnutt as a new chair; if DC is willing to serve, then he will replace Bob Kraft. Else Peter Robinson will chair. SH to inform Robert Kraft that he is being relieved of chair Due: no date given MSM to approach David Chesnutt and if necessary Peter Robinson Due: asap TR 3 Hypertext Members seem keen, and some work is being done with HyTime, but no work products have been produced yet. TR 4 Formulae etc. Met in January in Kentucky, but no minutes yet. TR 5 Newspapers Rieger will be here; MSM should speak with him. [This proved impos- sible. -MSM] RA asked what was wanted from this group. Layout, he said, seemed to be the only paramount issue; there are no constituent elements which we don't already have. LB objected that that was precisely what the WG should be deciding. AZ observed that some types of newspaper content are special: ads, classified, captions, comics. MSM asked whom the WG was to serve: different groups want rather different things. LB suggested three constituencies: linguists, industry, historians. MSM added that sociologists use newspapers for content analysis. The committee agreed that the first priority was to find someone to serve as head. The following plan was agreed on: 1. MSM to speak to Rieger. 2. If he won't or can't serve, get recommendations from him. 3. If the recommendations don't pan out, go back to old listserv logs and find the people who wanted newspapers included in the TEI guidelines. AZ asked whether this WG includes magazines and suggested including everything having to do with mass communications media: periodicals, newspapers, airline magazines. Possibly also scientific journals; no decision was reached. MSM to discuss work group with Rieger Due: asap TR 6 Language corpora No report from Biber. Anyone who sees him should ask. TR 7 Reference works No head. Status unchanged. TR 8 Printed Books LB has spoken with John Barnard and Jonquil Bevan. They have pro- posed Lotte Hellinga, an incunabulist. Bevan is a computer-aware 17th- century specialist, Barnard a respected 18th-century specialist, but computer naive. MSM noted a possible need for a further group on layout and typography, for the detailed transcription of sources, if TR8 did not address layout in detail. LB to discuss with Barnard and Bevan whether TR8 expects to treat Due: asap It was agreed to form a WG on presentation (layout, etc.), if TR8 does not expect to treat the subject. TR 9 Physical Description of Manuscripts, Paleography, and Codicology It was agreed that SH should contact J. Hamesse and discuss. Claus Huitfeldt to be given to this WG as a member, its scope to include mod- ern manuscripts. It was agreed that Elli Mylonas and C. Huitfeldt should be members of WG on MSS. A fourth member (historian) would be very good. Greenstein should be asked for recommendations; possibly also Elizabeth Brown. LB to consult with Greenstein re: historian for TR9 Due: asap MSM to consult with Elizabeth Brown re: historian for TR9 Due: after Greenstein's reply (and only if DG's recommendation does not work out) AI 1 Linguistics Good progress. AI 2 Spoken texts Nothing heard. LB has heard indirectly from Jane Edwards that SJ has contacted her. AI 3 Literature LB suggested that this work group's "Final Critique"(document AI3 W5) should be treated as a user response and comment. It has nothing to do with them as a work group. The minutes do need a response, in particu- lar: 1. Criticisms of TEI P1: * need original photocopies in examples * need attention to literary theory * need minimal length codes, minimal number of codes * LOB and OE examples should demonstrate entire set of tags * need documentation of existing schemes 2. Work groups needed: * WG on Literary Theory * WG on Performance * WG on Poetry * WG on Prose Texts * WG on Critical Editions 3. General TEI Procedures * TEI should produce software to work with Word Perfect and Word * New literary WGs to report to AI3, not to AI committee * AI3 to remain constituted until November 1991 The committee expressed some uneasiness with the implicit subdivision of the field; LB objected to it on the grounds that literary specializa- tion runs by period, not by text type: no one is a specialist in prose. MSM suggested that in fact many people identify their specialty with a language, a period, and a genre, and so the division is not a priori implausible. Many markup problems, he suggested, arise from peculiari- ties of the text type rather than from its language or period. SH proposed that someone should talk with PF and try to get some interaction, in order to forestall the apparent antagonism of the final critique document. If the views of the document represent those of many people, they must be addressed; otherwise the problem needs to be played down. It was agreed that SH and NI would attempt to clear the air with PF. RA suggested that AI3 was suffering from a systemic problem common to all the planning-level WGs: not enough has been said to the planning- level groups about what is to happens to them after they have proposed the formation of tag-level WGs. NI proposed that NI and SH speak to PF to get a step into the process of responding and to clear the air. So agreed. NI and SH to discuss with PF (1) the three areas addressed by the WG's comments, (2) the structural and organizational issues, (3) the complete absence of proposals for analytic tags, and the intended work of the WG on literary theory. Due: asap After discussion, lunch, and further discussion, the committee agreed upon the following responses: 1. Criticisms of TEI P1 to be treated as user response, as per LB's suggestion 2. Work groups suggested: * WG on Critical Editions has already been formed. Ian Lanca- shire and David Chesnutt to be drafted for service. (Charles Faulhaber to be used in MSS work group if needed, rather than in TC.) * WG on Performance Texts approved in principle, membership and reporting line (AI or TR) to be decided later. * WG on Poetry approved in principle, membership and reporting line (AI or TR) to be decided later. Verse tagging may have to be routed through both TR and AI. * WG on Literary Theory neither approved nor disapproved: more information to be requested. * WG on Prose Texts neither approved nor disapproved: more information to be requested. 3. General TEI Procedures * Software production remains outside the goals of the current TEI project. * The new WGs are to report to AI or TR, not to AI3 committee. (This was decided by a vote of 4:1.) * No funding allocated to a further meeting of AI3. AI 4 History This group has produced some internal papers surveying types of his- torical sources. Some detailed discussions, have been held, but not written up yet, on P1. Meeting in Southampton at AHC in April. AI 5 Machine-readable dictionaries Nicoletta Calzolari and RA had a planning meeting in January for Sat- urday's meeting. Agenda has been set: * Attempt to formulate a set of base tags * Attempt a set of grouping tags for which there might be no consen- sus. * errata and their correction * documentation of dictionaries (notes for TEI header), as, for exam- ple, in James L. Peterson, "Webster's Seventh New Collegiate Dic- tionary," 8 August 1986. * should there be a database standard (or: a mapping between SGML and DB representations)--i.e. should the TEI be talking about what goes into the database systems in which dictionaries will often be stored? Interconvertibility appears to be a requirement. Does that mean we should be defining a data model? (After discussion, it appeared that by "data model" was meant a rather informal treatment along the lines of the Amsler/Tompa paper, rather than a formal specification linked to a specific data model in the narrow sense. Solutions for all these issues were not expected in one day; individ- uals would be appointed to work on specific problems. This WG is now a tag-level work group. (Like any other, it may sug- gest need for further work groups.) RA and NC are joint heads. Members are: * RA (non-fundable) * NC * Carol van Ess-Dykema (non-fundable) * NI * Susan Warwick * Frank Wm. Tompa Also, a representative from the Consortium for Lexical Research (Louise Guthrie) will attend at least the Saturday meeting. Bob Ingria will attend for liaison whenever Susan Warwick can't come). LB asked the status of P1's chapter on dictionaries at this stage, and stressed that something concrete come out of the work group which can replace the relevant chapter, preferably with explicit DTDs. RA responded that it was one of several sources for the work underway. AI 6 Computational Lexica RA has received minutes of a meeting; NC met with Bob Ingria in Janu- ary. He had decided to form three subgroups to look at semantics, gram- mar (both generation and parsing), and machine translation. AI 7 Terminological Data This was approved as a new work group with Alan Melby as head. Further Work Groups It was agreed that we have enough work groups to go on with; there should be no active searching for further topics. Working Committees LB and MSM summarized the work of the Metalanguage meeting: papers and discussions on naming conventions, the issue of TEI conformance, SGML solutions to markup problems. The committee is taking a new approach to the ongoing problem of producing papers on sample transduc- tions. SH suggests mailing the naming conventions paper unsolicited to WG heads. MSM (through Wendy Plotkin) to mail ML W26 (naming conventions paper) unsolicited to all WG heads Due: upon completed revision of ML W26 The TEI conformance discussion was briefly described. Lynne Price has been certified as TEI liaison to ANSI X3V1 WG 8. LAP also working with DTB on TEI recommendations for SGML revision. RA recalled that the ML committee had been charged with telling us what the first step would be if SC chooses to make TEI a standard. It was agreed to renew the inquiry: what specifically would need to be done to make TEI recommendations a national or international standard? The Text Documentation committee is expected to meet during the 1991-92 academic year. 4 TEI GUIDELINES Discussion of comments on P1 was deferred until the next day. The committee discussed reprinting TEI P1. MSM suggested it would be better to wait for version 2, and believed that it would indeed make sense to release version 2. The form of version 2 was discussed. MSM described a possible split into 3 volumes: first, a text presentation like that of P1 but more direct, with less discussion and justification; second, a reference man- ual with DTDs, SGML declaration, formal pre-defined WSDs, and alphabeti- cal list of tags; third, a book of examples and interpretation (where the discussion and justification would go). A fourth document would also be desirable: a tutorial. The committee agreed that this split- ting of the Guidelines was worth further consideration. It might create problems for commercial publication -- but division into four parts needn't imply publication in four volumes. editors to draft brief publisher-oriented outline of multi-volume or multi-part form of guidelines, describing subject of book, description of function, and list of chapters. Due: next SC meeting (June 1991) 5 EDITORS' REPORTS ON OTHER DEVELOPMENTS None. Draft May 6, 1991 (18:17:02) TEI SCM17 Minutes, Tempe page 3 ------------------------------------------------------------------------ Part 2 FRIDAY 15 MARCH 6 AFFILIATED PROJECTS Goals and Interaction NI suggested that the committee merge NI's and MSM's proposals for affiliated projects and delegate to someone the task of actually naming TEI consultants from the list(s). NI, LB, MSM were so delegated. SH asked whether the number of APs could be reduced and what exactly they are to do when. NI replied that the APs are to encode sample text; the exact amount will vary depending on the project. LB proposed the following summary of TEI / AP interaction: 1. APs provide an URC 2. APs give us a description of their encoding scheme 3. TEI makes site visit with specific task of helping estimate resources needed for conversion 4. APs prepare a short sample encoding/conversion, as test 5. APs continue work undisturbed, working on larger sample encoding 6. second site visit to check on larger sample 7. APs produce report describing the difficulties encountered in applying the TEI scheme to their texts and how these problems were overcome From the APs, the TEI stands to gain confirmation that the TEI scheme is suitable for the work people are doing now, sample texts in TEI form, and possibly bug reports. It was unanimously agreed that work with the APs needs to get started immediately, in order that we can act upon any bug reports they provide. Various mechanisms for ensuring this were proposed. LB proposed setting a deadline, and scrapping the entire idea of the APs if there were not four APs working successfully with the guidelines by the deadline. SH proposed specifying a deadline of 15 June for the first report (to be the responsibility of the TEI consultant). Overall Funding Level for Work with Affiliated Projects NI asked how much money might be spent on the APs and suggested allo- cating a certain amount (e.g. $2000) per project. SH suggested the total expenditure should have a limit of $25K or $30K. LB noted that the following heads had been proposed for expenditures and asked which should be authorized: * workshop attendance * initial site visit * final site visit * honoraria * consultant workshop attendances * AP travel to WG meetings * software subsidies AZ and SH objected to authorizing specific amounts, on the grounds that the amounts needed to be seen in the context of the overall budget. Upon DW's motion, the committee agreed that the APs are very impor- tant, LB and SH specifying that they voted affirmatively only on the basis that we are committed publicly to the APs and should not make fools of ourselves. After discussion the committee authorized the expenditure of up to $30,000, but no more, for work with the affiliated projects. TEI Consultants and TEI Interaction with APs AZ asked whether enough people could be found qualified to serve as a consultant: they must be knowledgeable about the TEI, about SGML, and about the project's field of application. He suggested that a week-long meeting/workshop for projects might obviate the need for such consult- ants: the editors could lead the meeting and handle all the APs at once. Three models of interaction between the TEI and the APs were dis- cussed at length: 1. TEI consultants make one or two site visits to each AP ("site- visit" model) 2. TEI holds a single long meeting with the APs, and no consultants are named ("AP-workshop" model) 3. TEI holds a single long meeting with APs and consultants, follow- ing which the APs continue to work with the aid of their consult- ants ("mixed-workshop" model) The editors urged that naming consultants would help ensure wider dissemination of the TEI and broader training, as well as give the TEI a longer-term relationship with the projects; the workshop model was praised for its simplicity and for the opportunity it offered of getting our work with the APs done with quickly. AZ observed that since TEI consultants will provide answers to many questions in the APs, they would have to take great care to ensure that the same problem, arising in different places, received consistent answers. Upon SH's query, MSM estimated the cost of one workshop for both con- stultants and APs at $24,000: $ 9 000 AP per diems (15 projects * 5 days * $120 per diem) 3 000 consultant per diems (5 consultants * 5 days * $120) 12 000 airfare (15 + 5 airfares @ $600) $24 000 total Upon LB's motion to eliminate the AP-workshop model, RA and AZ asked to consider specific individual names of possible consultants, challeng- ing the editors to name even five possible qualified consultants. LB suggested several candidates qualified for training as TEI consultants (no claim made as to their willingness to serve): Steve DeRose, Susan Hockey, Stig Johansson, Elli Mylonas, Allen Renear, Harry Gaylord, Paul Ellison, Terry Langendoen, Gary Simons, Dan Greenstein, David Barnard, David Durand, Lynne Price, Frank Tompa, Nancy Ide, Bob Amsler. Upon MSM's motion, the mixed-workshop model was adopted, in the fol- lowing formulation: 1. A common workshop for consultants and APs to be held in May, June, or July. 2. One consultant to be responsible to each project. 3. Consultants to have continuing relationship with projects during production of large sample. 4. Consultants to receive $750 per project. 5. Honorarium contingent upon receipt of final report. 6. Consultants to be named by editors subject to 5-day veto by steer- ing committee (AZ to be sent fax). In a ruling from the chair, DW observes that under this rule four negative responses will be needed for veto. A motion to amend to allow veto by individual members of the steering committee fails. The proposal was approved with 4 affirmative votes. 7 FINANCIAL ISSUES The budget statements provided by MSM were reviewed, and a number of inexactitudes found. The amounts shown for travel by Greenstein and Ellison were suspicious, and AZ's Chicago trip of July 1990 should be listed under the first cycle, not the second. [The Greenstein and Elli- son amounts were discovered upon inquiry to be clerical errors: the actual amounts were about 100 ECU each, rather than 500.] DW suggested a more consistent column order was needed in the reports. SH praised the work that had gone into the reports but still found them hard to read and asked how much money was actually left. MSM replied, working from page 1.2 of the report: ECU expenses 13 361 ECU committed 44 244 total spent/committed 57 605 Oslo 10 000 Oxford 11 856 total really committed 79 461 EEC contract 90 000 uncommitted 10 539 NI reported that the ACH council should approve the opening of a bank account for TEI funds in Chicago at their meeting in Tempe. The commit- tee approved the payment of future reimbursements in the traveler's local currency rather than in ECU, even though this may incur signifi- cant bank charges. 8 COMMENTS ON TEI P1 MSM summarized the comments on P1 received so far, focusing on the general issues. The major issues: * verbosity vs. brevity * simplicity vs. complexity (information-rich and -poor) * software support 9 FUTURE STEERING COMMITTEE MEETINGS The next meeting of the Steering Committee was set for Sunday and Monday, 16-17 June 1991, immediately preceding the ACL conference in Berkeley, CA. 10 ESTABLISHING AN ENDURING STATUS FOR THE TEI DW summarized the situation thus: establishing an entity is easy; finding the funds is the hard part. Settling the purpose is necessary too. DW suggested the TEI might be set up as a consortium; the TEI may be recognized as valuable for R/P center, Linguistic Data Consortium, NEC etc., and may possibly get money from all them. We must not only ensure the TEI's institutional survival; we must ensure that extensions and revisions can be done, publications distrib- uted, and software developed. AZ said the basic question was what must be done in June 1992. SH will put this on the June agenda and draft a brief paper on the issues to be considered. SH to write a paper on the requirements for the future Due: circulate by 1 June 1991. By 15 May, SH would like any comments or suggestions. 11 USABILITY/SOFTWARE RA asked whether the TEI has the resources to produce software, and observed that we are committed to the production of documentation, but not to producing software. SH noted that the usability issue may possi- bly be an information problem: there is a perception that the TEI scheme isn't easy to use. The workshop may help to dispel this notion. RA noted that easy-to-use SGML software does exist (e.g. Arbortext Pub- lisher, with SQL interface, etc.); the problem is price: it's out of the league of the low-budget humanist. MSM suggested that we encourage development of TEI software by releasing a formal grammatical description of the TEI subset of SGML. Approved, ML committee to be responsible. Metalanguage committee to draft and release a formal grammatical description of TEI documents Due: no date specified MSM suggested further that the steering committee authorize $5000 for software purchases on this basis: if any SGML-conforming parser which parses, normalizes, and validates, is offered to us for under $100 per copy, we will buy up to ten. This was approved. 12 ADJOURNMENT The meeting adjourned at 6 p.m. Friday 15 March 1991. Draft May 6, 1991 (18:17:02)