Minutes from November 2, 2011

TEI in Libraries SIG meeting

(adjunct to Digital Library Federation Fall Forum, Baltimore, MD, 2 November 2011)

At a Baltimore hotel society gathering... -- B. Dylan

Attending

Dave Lacy (Villanova), Melanie Schlosser (OSU), Kevin Hawkins (UMich) (convener), Glen Worthey (Stanford) (note-taker), Dot Porter (IU), Perry Willet (CDL)

TEI in Libraries "Best Practices" update

1. v. 3.0, October 2011, of Best Practices document completed and launched at 2011 TEI meeting in Würzburg.
2. Next steps: Kevin to present BP ODD to TEI-C Council for long-term institutional maintenance as a TEI Customization. TEI Council meets next week; Kevin plans to make pitch for this transfer of stewardship for the document from our SIG to the Council.
3. TEI-Tite
   1. Description (refresher): vendor spec designed esp. for the AccessTEI program, currently administered by Apex CoVantage. Related to, but separate from, BP work.
   2. Some issues have occurred during Tite implementation process: Apex has made a few changes (manual edits of DTD) that weren't folded back into Tite; Kevin has been gathering records of these changes, integrating them into the Tite ODD, collecting bug reports and tickets to make sure all these changes are accounted for.
   3. Kevin has edited both Tite and BP appendices, wherein the two documents refer to each other, to better clarify their relationship.
4. Google Books news
   1. Kevin has been assisting Google engineers who've expressed interest in producing TEI output from Google scanning of library collections. All agree that this is big news, and great news!
   2. He received communication about this through at least two different channels: directly from Martin Mueller and from emails to Google's library partners.
   3. Kevin has convinced Google to use the BP document as reference, to make their output conform roughly at "arbitrary" encoding levels (i.e., it's likely their output would be at level 3-4, but definitely should be encouraged not to throw anything away for the sake of "conformance" to a particular level, e.g., any structural information they may have garnered in their automatic recognition processes)
   4. Some discussion of the importance of this development, if it pans out.
5. The "Interoperability" question _(This came up at the end of the meeting, but fits thematically with discussion of the BP document and the TEI Guidelines)_
   1. Kevin responded to general complaints ("the September events") that TEI hasn't done enough to enable interoperability and interchange of texts
   2. Kevin has discussed with unspecified others a proposal to revise the Guidelines to more honestly address what the TEI actually does -- to weaken the claims of "interoperability and interchange" -- or even simply to stop claiming that this is an overarching goal of the TEI. This revision could mainly restricted to prose descriptions, rather than some attempt at "fixing" the encoding standards. That group plans to formulate revisions and share for wider comment.

Possible future directions for the TEI in Libraries SIG

1. Enhance BP document with better options for encoding of serials
2. Create XSLT for converting Tite to conformant P5
3. Enhance the TEI's relationship with various current library standards and practices, including:
   1. Clarify TEI's relationship to FRBR. (It would be good if BP, or TEI as a whole, had a way of explicitly encoding which FRBR Group 1 entity is the object of encoding. Kevin believes this could be problematic, but worth investigating.
   2. Perry expressed related interest in providing mechanisms for linking to LC authority files and other outside metadata such as id.loc.gov. Kevin reviewed proposed solutions from the past in this space.
   3. Dave suggested using "mdref" from METS or similar; Kevin supports this idea, even proposing using the METS namespace to do so.
   4. Dot noted importance of strong library interest in Linked Data / RDA, and all agreed on desirability at least to address this, along with crosswalking to other metadata formats besides AACR2 (which is what we limited ourselves to in the current BP document). Kevin suggested that addressing this could be as simple as adding RDA as one of the examples of a national cataloging code (along with AACR2). If MARC is superseded, we would need to revise our field mappings.
   5. Should we have our own recommended METS profile? No consensus.
4. Interest expressed in Würzburg for registry of Tite and BP implementations so that people engaging in new projects don't have to start from scratch. This could be done organically and informally, e.g., in the TEI wiki, etc.
5. Create wiki page bringing together TEI customizations designed for interoperability, including TEI Analytics, Text Grid's "baseline encoding," and the BP document (and maybe Abbot as well)
6. The SIG's future mission could focus more on evangelization for use of TEI in libraries. The BP document was really all about this, but even now that it's "done," the need for evangelizing remains.
7. The SIG could also engage in fostering this TEI-library relationship in the other direction, viz., representing the library community to the TEI Board / Council. This has happened fairly effectively so far through TEI Council service by several "library" people (e.g., Kevin, John Walsh, others)

Existential question: the SIG Itself

1. The current leadership (Kevin and Michelle) has expressed an interest in passing the baton to others. It was suggested that the TEILIB-L list may be an appropriate venue to solicit interest in the SIG. Although this list has not been very active in recent years, Kevin reported that it has about 80 members. (Kevin also says that there's a forthcoming article in the Journal of the Text Encoding Initiative that includes membership totals for the SIG mailing lists.)
2. We discussed the relationship of our SIG to the many others (see http://www.tei-c.org/Activities/SIG/). It was generally agreed that there was some overlap (e.g., with Scholarly Editing, Ontologies, Education), but not really that much.
3. Although it seems that the work of our SIG has reached some conclusions, and although the TEI Council believes that it's okay for SIGs to come and go as their work finishes, we didn't feel like our work in libraries is necessarily complete, nor is anyone ready to dissolve it. (This in spite of general acknowledgment of the impression that TEI adoption in libraries is flat at best, or even declining -- somewhat ironically, given potential Google Books interest. Perry reported, e.g., that TEI is almost not used in UC system; Mark Twain papers last project to use TEI.)

Conclusions

This meeting was short on actual resolutions, but discussion of the above issues will continue in the wiki and TEILIB-L.