TEI Council Meeting

All times are local (Central European Summer Time, UTC +2) unless otherwise noted.

In attendance, from TEI Council: Laurent Romary (LR; chair), Peter Boot (PB), Arianna Ciula (AC), James Cummings (JC), Daniel O’Donnell (DO; Board Chair), Dot Porter (DP), Sebastian Rahtz (SR), Paul Schaffner (PS), and David Sewell (DS; minutes). Representing Oxford editorial support group: Lou Burnard (LB). Not present: Gabriel Bodard, Elena Pierazzo, Manfred Thaller.

The meetings were held at l’Ecole normale supérieure Lettres et sciences humaines, Lyon, France. Most Council members were also present for the Public Meeting held at ENS on April 1, the day prior to the Council meeting, and several served as session chairs. The TEI Council wish to thank our hosts at ENS-LSH for their hospitality, and acknowledge in particular Serge Heiden for going beyond the call of duty in arranging back-up facilities during the Thursday jour de grève at ENS.

The meeting was called to order on Thursday at approximately 09:10 and continued until approximately 18:30; it resumed on Friday at approximately 09:00 and continued until approximately 16:00.

Discussion under this heading focused on two topics: general organization and workflow of the TEI website; and maintenance of the TEI “vault”, i.e. historical versions of the Guidelines and archived versions of releases going forward. LR summarized the four main issues:

• updating our tools
• updating files
• keeping the vault in a stable place
• reorganizing the website

Question: should there be a formal facility to allow user comments on the Guidelines (primarily or only the element/attribute/class reference sections), along the lines of the “user contributed notes” feature on www.php.net? The purpose would be to allow users to provide examples of how they use TEI features, notes on best-practice usage, etc. Council agrees that (unlike the case on php.net) this should not be done within reference pages themselves, but rather via links to dedicated areas on the TEI Wiki. This can easily be done by batch-creating template pages on the Wiki, and having the Guidelines stylesheet generate links to them for each reference page. The only tricky technical issue is whether we can easily differentiate links to Wiki pages that have comments versus ones that are empty.

SR and JC will investigate implementation and report back to Council

This item was an update on the status of TEI Tite.

There is some urgency over finalizing the recommended format of the TEI Tite customization, as the TEI Consortium received Mellon funding with a deadline to put out an RFP to conversion vendors interested in bidding on bulk conversion work using the Tite specifications. However, there is some confusion over how final recommendations for change have been gathered and implemented. The TEI in Libraries SIG has been working on this and forwarding suggestions to Perry Trolard, but PT is not able to implement them directly on Sourceforge.

Council agreed that as Tite is a formally recommended TEI customization, Council has technical authority over it. We need to vet and approve changes, then commit them to SourceForge.

There was some discussion and disagreement over whether Tite should be seen as a static object or as a basis for customization. Consensus was that (1) Tite as formally recommended by the TEI will be static within the context of a given contractual agreement between content owners and conversion vendors (as in the main RFP), but (2) individual projects may wish to customize Tite for their own purposes, and that is something we encourage.

DO will issue a final call on TEI-L for comments on Tite. Council formally asserts control over Tite, and DO will ask PT to collect proposed changes to the existing Tite customization and forward them to Council for approval.

LR presents this proposal, which is to approve a new Special Interest Group on scientific bibliographies with the goal of providing recommendations for the use of biblStruct for bibliographies in scientific/technical literature. LR notes that people are already using TEI for scientific publishing, so we need to put forth best practice recommendations and possibly schema modifications. The strategy is to demonstrate the usefulness of biblStruct as a way of demonstrating the power of TEI to other communities, as the XML format for these document structures. Council agrees on the utility of doing this; AC notes that we’ll want to provide a specific tutorial on this use case.

APPROVED. As LR is already part of the proposed SIG, LB volunteers to be the official Council liaison.

Elena Pierazzo (in absentia) presents a report on activities of the MS SIG. They have proposed some tweaks to manuscript elements, and will have a workshop on genetic editions in May 2009.

No action item, just be aware SIG is active.

LB notes that two different communities in France are interested in working on improvements to Guidelines in this area. LR thinks that most people are quite happy w the current state of the Guidelines chapter, but there is a need for more work on marking synchronization with audio files. LB adds that a separate German group also feels there is serious need for revision. [action item??]

DO reports on the grants approved this year to provide funding for SIGs to support activities toward defined goals; Susan Schreibman issued a call for applications in March 2009. Two Council members, JC and DS, are serving on the committee that will evaluate and approve requests. This is an experimental program; if it works, the Board hopes to expand it.

Some discussion follows on other types of TEI funding. DO notes that funding is in principal available for any chartered TEI group, e.g. working groups. However, as LR points out, Council needs to have supervision over the funding of anything related to the TEI technical mission. Any grant involving tools, software, or the Guidelines is Council-related.

SR reported on his work with the new Vesta release, and the current transition to XSLT 2.0 stylesheets for schema generation. (In the aftermath of Council discussion in Lyon and on tei-council, SR formally released to TEI-L the roadmap / project plan for Roma and ODD processing based on Council agreement about our goals. The following paragraph briefly summarizes the Lyon portion of that discussion.)

Basic issue: once work began on Vesta, it was necessary to fork the XSLT stylesheets to XSLT, for use with Saxon in a Java environment. But Roma runs on libxslt, which supports only XSLT 1.0 [and which may remain there indefinitely—DS]. So: do we want to maintain the forked stylesheets, or are we willing to commit our ongoing resources to Vesta’s XSLT 2 environment, with the goal of an eventual rewrite of Roma?

Council agrees it is undesirable to maintain more than one set of stylesheets. On the other hand, we do want to retain OpenOffice transforms as a TEI deliverable. SR proposes that we call WebRoma the ODD editor; Vesta has a different function. We should also make a distinction between TEI services and interface [one service might be instantiated in more than one interface]. DO suggests that the strategic direction as seen from our users’ perspective is: there should be single locus for all TEI-related services; the technical backend should be opaque. LR: We implement it as an SOA (service-oriented architecture), meaning that we identify the processors that are behind the scene as services. DO: output services (in the non-technical sense) need to be bundled together, but not necessarily input services. The set of people who actually author and compile ODDs is much smaller than people who use TEI and generate output.

SR will come up with a roadmap for the Roma family, propose it to Council, then disseminate it on TEI-L. [DONE as of 11 April].

Most of the discussion focused on submitting a proposal to the European Science Foundation (ESF) for one of their Exploratory Workshops.

LB suggests that we could use a workshop as a way of beginning the planning for P6; we could also focus on the role of TEI with respect to various communities of research. AC says there is funding for science-driven research workshops (science in the sense of Wissenschaft) with a maximum 30 participants (who must be specified in advance), lasting 1-3 days. Deadline for proposals for the current grant cycle 30 April. The location is flexible. Requirements include wide representation from across Europe, and no more than five participants from the proposing country. Maximum funding available is €15K. AC feels that participants should be people who will be involved in future TEI development.

Strasbourg would be a possible host site; LR has access to space in the Tour Montparnasse, so Paris would be another. We might also investigate cosponsorship, e.g. with Adonis.

LB and LR will draft a workshop proposal. AC will not be involved in writing the proposal, to avoid conflict of interest.

(As secretary was participating in discussions on these items, notes are sketchy. See also any comments on SourceForge.)

2209933 content model of am

• Argument from use cases. Cannot put unclear etc. outside the am if you’re using diplomatic transcription. LB objects that am includes a marker of abbreviation, seen on page but not used in transcription. Prototypical abbreviation is ‘.’. But use case is: 3-letter abbreviation marker where one is unclear. JC: let’s modify documentation so people don’t abuse this. We agree that usefulness outweighs the possibility of abuse.

  APPROVED

2493417idno coverage.

• We agree on the simple application + corresponding change in the prose. Change definition of idno to say it can be used to identify any part of a bibliographical description; 2nd, add it as a possible content model for author. LB/SR are not happy about idno as a child of author. LR suggests that in the scientific world we want to be able to create bibliographical entries that include idno‘s that connect to our data, e.g. standard number for publication or author. LB: there are two solutions. Either we say we’ve changed our mind about idno and don’t restrict; or we restrict the places where it is allowed. JC suggests: why can’t we use ident for this purpose as it is already allowed? … maybe we need a special bibliographical phrase class (in order to get ident inside author). There are various other elements inside a bibliography that can take identifiers. Consensus: we will add a bibl.Phrase model class, which will contain idno.

  APPROVED

2517629 Guidelines: type and resp in note

• OK, accept as proposed.

  APPROVED

2607768figure should be global

• Rationale: graphic when not inline should be wrapped in a figure as best practice. Accept as proposed.

  APPROVED

2673045 new attribute rational on num

• This was discussed in detail on tei-council and SF. The majority voted for adding valueType. Consensus: we implement, with a closed value list: “decimal”, “float”, “ratio”.

  APPROVED

2672530 Remove repository element requirement in msIdentifier

• Fine; we will add to documentation of repository that it is not meant to indicate just any geographical location.

  APPROVED

2242434 element for text suppressed by the editor

• The proposal was for an element like suppressed with a function opposite to that of supplied. After discussion, council comes around to point of view that yes, supplied does perhaps require symmetry. But suppress is not the best name; how about omit. This is semantically more neutral. Council does not totally agree that omit is needed when sic exists, but in the interest of flexibility, we can add it. We need actual use case examples.

  DEFERRED: we will ask the proposers for actual use cases (and in particular, cases where typed sic would not be sufficient).

1954920 (role to be returned to att.naming)

• Agreed, add role to att.naming. [As LB notes in SF ticket, att.naming needs reorganization].

  APPROVED.

1925127: Provide discussion of taxonomical names

• Council sense is that it’s too complex. LR: is it really role of TEI to figure this out? Or solicit advice from specialists? We don’t have a proper TEI tool to represent full taxonomy. Suggest we leave open for now. No one has come up with a strong proposal. We need stronger set of use cases.

  DENIED: close ticket

1933481: ODD needs a way to indicate deprecation at next major release

• LR suggests that we simply add a status attribute in ODD to elementSpec or attributeSpec. Discussion was inconclusive; the whole issue ties in with larger discussion about ODD format.

  DEFER: leave ticket open.

1962972: new members for model.quoteLike

• PB explains the context. LB thinks there’s a good case to be made for using classes. All agree.

  APPROVED

1999652: Add who to quote

• LB thinks this is just a resurgence of basic problem that people mistakenly believe q and quote are the same. But when quote is rightly understood as part of a bibliographic construct with cit available, a new attribute isn’t needed.

  DENIED

2055891: Placement of Schematron rules

• SR summarizes his proposal to add constraint as a child of schemaSpec, elementSpec, attDef. Consensus is that we should implement; SR will come up with details before the next June release, in conjunction with Council. [See also below, Friday discussion; and discussion on tei-council, first week in April].

  APPROVED pending implementation

2525586: scope attribute on rendition/ for CSS pseudo-elements

• Was discussed in some detail via email. Accept. LB will incorporate DS’s response in the SF ticket.

  APPROVED

2531384: Rationalise application of target

• DP presents. She likes Lou’s suggestions. After discussion, decide we have to keep target so as not to break things. Define a class to define target values as one or more tokens; in cases where there must be only one, use Schematron. What about cRef? Leave it alone. Lou’s ticket note: make target point to 1:n, and define as class; enforce in Schematron.

  ACCEPTED

2594174: dimensions needs loosening up slightly

• Per Lou’s resolution on SF ticket: Added new dim element, and class model.dimLike. Redefined dimensions content as ( dim*, model.dimLike_sequenceOptional). Updated discussion in MS accordingly. Updated at rev 5985.

  ACCEPTED

The document has not progressed as far as we wanted; 7 of 13 sections are drafted and ready for comments. PB proposes that if no one else volunteers, he will write the remainder. DS asks whether we should go outside Council for authors; PB thinks not. LB: do we want to internationalize the translation? JC/LR: we can think about it, don’t want to have anyone start on it.

There is room for TEI to become better known in standards communities. There are various European agencies that have an interest in TEI; we should pursue our roles in them. On a higher level, we could promote TEI by giving funders more info on TEI, TEI best practices. (DO: this was mentioned by funders at the 2007 Members Meeting.) DO: we should be doing this, and also identifying agencies that are not requiring TEI, and explaining the benefits to them. But we are not in a position to do TEI certification. We’re not an agency, don’t have paid staff that can do this.

DO suggests that we prepare a package on digital resources and the TEI to share with agencies, and offer to give workshops. The TEI can profile itself as a meta-research infrastructure. … We should be a research NGO. [DS nominates this as the Quote of the Meeting.]

AC, DO, LR [with JC and Julia Flanders]: talk further about drafting communications to outside agencies (approval for these will have to come from Board however)

JC: As a follow-up from our Galway meeting, we might want to come up with a framework for indexing/retrieving TEI examples in the Guidelines (or even external ones). We want to be able from a reference page to be able to point to uses of examples. All examples in Guidelines should have xml:id and xml:lang. [SR feels the latter is unneeded because xml:lang is inherited. But LR feels that we should decouple the language of examples and the language of the enveloping Guidelines for modularity. JC notes that there are some examples that incorrectly inherit xml:lang=”en”]. LR suggestion we might move away from inherited examples in XML to a library stored outside and pointed to, carrying xml:lang. JC proposes decoupling examples from Guidelines and making them referenceable.

Counterarguments: (SR) this is not our most important task right now; (LB) decontextualized examples don’t always make sense. (SR): The bigger problem is that examples we have are too small. LR: so we need some kind of way to externalize examples. DO: The first thing to do is to file a bug report on incorrect xml:lang on examples; then add a feature request to add xml:lang to all egXMl examples.

Council agrees it would be good to have an index to examples in the HTML format as well as in PDF. Also, a “click on show me other examples” link, to show all (?) available examples. [Since implemented by SR.]

SR points out that our HTML rendering of Guidelines is just one possible rendering; the reason we make our source open is so people can build on it if they want. So let’s separate out HTML presentation from the collection of more examples.

DS will gather a list of examples with incorrect (explicit or inherited) xml:lang values; JC, submit feature request to add xml:lang to all egXML elements.

The first stage of Oxford’s work for the ISO was to show that TEI could describe an ISO standard; the second stage was lossless bidirectional translation between TEI and MS Word. This was achieved and demo’ed to standards bodies. The proposal then is to go on to a pilot project to implement some of the standards. (LR asks if there wasn’t a standard prepared in XML recently; SR: yes, but not validated by ISO.) There were objections in stage 1 about the TEI table model (probably CALS tables would need to be used instead) and questions about math. LR notes that Vesta supports an ISO-looking document (but not ISO-approved, as SR notes). SR isn’t sure that this project will go any further; Council should perhaps not spend more time on it right now. LR: can we say that the experiment shows that a subset of TEI is an appropriate tool for the task? SR: yes, because we understand the necessary structure and terminology, plus have the ability to customize. And the MS Word stuff demonstrates that there is a stylesheet framework available Word-TEI adequate for authoring.