Minutes of the Meeting Held at Oxford University Computing Service, November 17-18, 1990 Document Number: TEI ML M33 Draft November 21, 1990 (12:41:21) Lou Burnard Present: David Barnard (DB), chair; Lou Burnard (LB); David Durand (DD)Jean-Pierre Gaspart (JPG); Lynne Price (LAP); Michael Sperberg- McQueen (MSM); Frank Tompa (FT). AGENDA The editors requested the addition of two further items to the agenda previously circulated by DB as MLA29: these were items 12 (Discussion of DTDs) and 13 (Discussion of initial reactions to the Guidelines). 1. CURRENT STATUS OF THE TEI The editors reported that about 500 copies of the first draft propo- sals (TEI P1) had been distributed. DTDs had been tested and were to be placed on the fileserver shortly. A Workshop had been held in Chicago, with success directly proportionate to the enthusiasm of its partici- pants. A European Workshop would be held in the New Year. MSM briefly summarised the TEI committee structure during phase 2 of the project: the Metalanguage and Documentation committees would continue as before. As Text representation and Analysis committees began to consider more abstruse or specialist areas, plenary discussions would be replaced by smaller more focused work groups, with a typical shelf life of 3/4 months and a typical membership of two to three. Fifteen such groups were currently in the process of being set up, some to work in a direct- ed way to produce specific tag proposals for a given area, others to survey an area not yet addressed by the Guidelines. DB had given a status report at SGML90, which LAP reported had been well received. She also noted that Anders Berglund had specifically asked for TEI comment on his revised version of ISO TR 9573; the editors agreed to pass this information to the head of the Character Set work- group. LB To ask Gaylord to contact Berglund Due: Not specified JPG gave a brief outline of interesting developments in the use of SGML to provide a formal description for the 'directors script' in mul- timedia products such as CD-I. It was agreed that this would be of rele- vance to the Hypertext workgroup. MSM To ask DeRose to contact JPG Due: Not specified 2. ADDRESS LIST The committee address list was circulated and amended. A new version is attached to these minutes. 3. MINUTES OF THE PRECEDING MEETING These were adopted. 4. MATTERS ARISING 4.1. DB to report on SDIF. A brief working paper (ML W24) had been produced, stating that SDIF would be usable for TEI purposes, if only there were any software which supported it. It was noted that the bundling of entity sets (etc) was a different problem from that of mapping character sets for transmission, and that SDIF addressed only the former. MSM asked for clarification on ASN.1 and whether there was any documentation for it more comprehensible than the standard itself. 4.2. MSM to revise ML W18. DB asked whether this document would continue to be relevant to subse- quent versions of the Guidelines. It was agreed that there was a contin- uing need for presentation of the theoretical issues underlying the group's recommendations. MSM To revise ML W18 in line with comments made at Chicago meeting Due: 15 Feb 91 4.3. LAP/JPG to draft Parser Pitfalls Document (W25) LAP circulated hardcopy of a document 'Diary of a Parser' outlining some of the problems. She also had some tutorial notes for which permission was needed before they could be incorporated. LAP To send Wendy Plotkin a draft cover note and whatever informa- tion was needed to seek copyright permission to include tutorial notes Due: 15 Dec 90 4.4. JPG to draft Minimisation document JPG presented apologies. A sample document addressing tabular matter was in progress. JPG To draft document on minimisation (ML W22) Due: 26 Nov 90 4.5. DD/MSM to draft document on naming conventions etc. DD reported that the document was progressing. DD To draft document on naming conventions (ML W26) Due: 26 Nov 90 4.6. NI/DB to draft section 9.1 of TEI P1. This had been done and incorporated successfully into the Guidelines. 4.7. Transduction Workgroup This had met after the Chicago meeting and work was in progress. It was agreed that this should result in a new document (MLW30) with a view to including it as an Appendix to the Guidelines. JPG had already done some work on the basic primitives needed for a formatter in his RTF to SGML transducer and would send details of these to DB. DB/NI To draft MLW30: Principles and examples of transduction to and from TEI format Due: 1 Jan 90 5. DOCUMENT REVIEW The document register was reviewed. It was noted that not all non-paper documents were as yet on the server. DB To ensure that all non-paper documents were placed on the TEI-L server Due: 26 Nov 90 The following documents were retired: W13, A21, W27. The following doc- uments are being revised: W14, W17, W18, W19. W25, W26. W28 (list of SGML vendors) was circulated at the meeting. DD To revise ML W17 (Housestyle) incorporating comments made at Chicago meeting Due: 1 Jan 90 6. AOB None was presented. 7. ISO LIAISON DB noted that ISO 8879 was coming up for its five year review, and suggested that several members of the committee were well-placed to influence this. JPG commented that personal comments were useful but more influential when documents had been formally introduced. DB asked what specific points we wished to convey (for example, that the two functions of SHORTTAG should be separated). DD requested clarification of committee members' formal links with ISO, which was provided: DB is a member of the official Canadian shadow group for SC18/WG8; LAP an occa- sional participant in the equivalent US group (ANSI X3V1/TG8); JPG expects to be re-accredited as Belgian representative on SC18/WG8 when his company completes its first year. There was some discussion of the ODA standard, which we were commit- ted by the EEC contract to investigate and which had considerable stra- tegic importance in the European context. JPG referred to an 'Express Project' which had abandoned ODA. MSM To provide check list of the information required by EEC con- tract Due: 31 Dec 90 FWT To provide information on how ODA documentation can be obtained Due: 26 Nov 90 DD To draft initial report on relevance of ODA to TEI (allocated document number W31) Due: next meeting After some discussion of whether we wished to register DTDs with ISO, three general points were raised on which TEI might produce input to SGML2. These were: * The standard should be a technical specification rather than a piece of legislation * Use of CONCUR, specifically: - it should be possible to support different views, in some of which data can be excluded - views should be processable in parallel, not sequentially - arbitrarily many arbitrarily nesting hierarchies should be sup- ported * LINK needs clarification It was agreed that a working paper should be drafted covering these and similar concerns as an official TEI proposal to WG8, in time for its next meeting in May 1990. DB to propose to the steering committee that LAP act as official TEI liaison with ANSI X3V1 Due: 26 Nov 90 DB To draft Notes for proposed revisions to ISO 8879 (allocated doc number W32) Due: 1 Jan 91 MSM asked whether the Guidelines should also be officially transmit- ted to ISO. After discussion, it was agreed that only a relevant subset of TEI material should be sent, to demonstrate the kinds of applications involved, and the rationale behind our proposals. These would include specifically W32, W18, Appendixes A & B and chapter 3.2 of TEI P1. DB To package a suitable set of documents for transmission by the TEI Steering Committee to ANSI X3V1 Due: 28 Feb 91 8. POLICY REVIEW: INTERCHANGE RESTRICTIONS Noting that among the concerns most frequently voiced about the usabili- ty of the Guidelines were their perceived verbosity, the committee reviewed its stated policy on use of minimization, name lengths etc. JPG pointed out that storage costs were now a quarter of what they had been when the recommendations were first made. Some proposed abbreviatory conventions could be parsed directly by using DATATAG or SHORTREF. This remained a matter for local processing. There was a need for an overall TEI processing model. After much discussion, it was agreed that further guidance on usabil- ity issues was needed. As well as a hand-holding tutorial, a collection of hints and tips was needed, possibly including some of the theology, some example filters etc. This lead into a discussion of TEI conformance and TEI validation. TEI-conformant interchange should be possible with- out alteration of the TEI SGML Declaration, but might require some addi- tional processing to convert into the ISO 646 Subset. JPG argued that conformance to the TEI DTDs necessarily implied support for concur; DB that subsets might be defined. It was noted that we could not control problems of physical character interchange, unlike e.g. NAMELEN which arose only after interchange. Was a system conformant which saw only the base doctype, provided that its limitations were documented? The topic of conformance was deferred. There was further discussion of how far the TEI usage of SGML should be simplified for simple-minded processors. A consensus was reached that SHORTREF, SHORTTAG and LINK should all be precluded by the TEI SGML declaration. The discussion of SHORTTAG in TEI P1 needed to be reformulated. During subsequent discus- sion of SHORTTAG it was noted that simplifying SGML usage also simpli- fied intelligent compression of documents for transmission. MSM To revise SGML Declaration, 2.2.2 in TEI P1, and MLW13 accord- ingly Due: 26 Nov 90 9. POLICY REVIEW: DATA CAPTURE ISSUES MLW22 would address these more specifically. A need for transduction tips and techniques was identified. There was some discussion of the advisability of providing some example application-specific filters e.g. RTF to TEI, Nota Bene macros etc. FWT felt that people were unconvinced by general cases; an JPG that it was important to convey the underlying principles; DD pointed out the immense variation in sophistication amongst the TEI community. People should be encouraged to place individ- ual solutions on the fileserver. LB noted that several transductions for a common text would probably be prepared for the European Workshop. It was agreed that more tutorial material on transduction techniques was required. MSM noted that in addition the material on extension tech- niques in TEI P1 chap 8 needed to be rewritten from the user's point of view. JPG proposed that transduction should be described in an applica- tion independent manner; rather than invent another metalanguage, it was agreed that plain if formalised English would be used. After some brief discussion, FWT's suggestion that a few examples be attempted before formalising this descriptive language was adopted. A need was identified for advice on data capture from scratch (i.e. without any pre-existing format), using for example style sheets, key- board macros etc. JPG remarked that LINK was a "foetal stage" in SGML's standardisation of the style sheet mechanism, with the added benefit of context sensitivity. It was agreed that this provided a useful conceptu- al framework for describing both data reclamation and data capture. all To reconsider MLW30 in the light of experience Due: March 1991 11. TEI SOFTWARE Several attempts were made to describe the processing environment and the place of TEI in it, and hence the kinds of software needed to sup- port it, but no consensus was achieved. DD noted that TEI provided both an abstract model which local formats might reflect and also a partial implementation of that data model. Insisting on SGML as the 'pivot' (that is, making the canonization process transparent) was difficult to justify in the current absence of software. The simpler that canoniza- tion software needed to be, the better. MSM expressed the need for specific DTD handling software tools with- in the project. The editors needed to generate clean DTDs, reference lists and indirectly-named (pace chapter 8) DTDs from a common base. There was some discussion of the best way of storing and processing this base: whether as an SGML document, perhaps defined with a different con- crete syntax, or as a relational database. A DTD tool capable of turning attributes into GIs (a tomperer) was also requested. Editors To review their requirements for DTD tools Due: Unspecified DB To investigate production of a tool for generation of indirectly-named DTDs Due: Unspecified MSM reported on discussions held earlier with Rich Giordano, about possible development of TEI-related software within the University of Manchester's Computer Science department. LB reported on other on-going software developments known to him. It was agreed that the TEI should try wherever possible to encourage such development without however endorsing any particular products. A requirement for some way of informally assessing software was iden- tified. A checklist of features could be provided, together with a bag of test applications, as a first step to formal conformance testing. It was not clear whether this should be provided by the ML committee or by some other work group. LB To formulate proposal for software assessment to be passed to Steering Committee Due: Unspecified 12. DRAFT DTDS MSM requested comments on the layout, style and general organising prin- ciples of the draft DTDs as soon as possible. all Feedback on draft DTDs to Editors Due: 10 Dec 90 13. COMMENTS LB described briefly how comments and responses were currently being organised: the editors would split detailed comments into briefly sum- marisable points. It was agreed that bundled tags on which ML committee input was needed would be posted on TEI-META, and that DB would co- ordinate a response within three weeks. Such postings would be identi- fiable by a subject line in the form 'URC nn - please respond by due- date' The committee then discussed some points raised by such of the com- ments as were to hand. 14. NEXT MEETING This would probably be held in Toronto at a venue to be determined by DB, probably between 4th and 8th March 1991. Draft November 21, 1990 (12:41:21)