Minutes from April 28, 2008

TEI in Libraries: Home

Meeting Date: Monday, April 28th, 8:30 am - 12:30 pm Note taker: Lisa McAulay, 4/26/2008 Notes edited by: Michelle Dalmau, 6/4/2008

ATTENDEES

 * Syd Bauman, (Women's Writing Project, Brown University)
 * Glenn Worthy, (Humanities Digital Information Service (Library), Stanford University)
 * Perry Trolard (Humanities Digital Workshop (Arts & Sciences), Washington University)
 * Melanie Schlosser, (Library, Ohio State University)
 * Jenn Riley, (Digital Library Program, Indiana University) * Chris Powell, (University of Michigan, Digital Text Services)
 * Michelle Dalmau (Digital Library Program, Indiana University)
 * Matt Gibson (UVA)
 * Perry Willett (Digital Library Publication Service, University of Michigan)
 * Natasha Smith (CDLA, UNC-Chapel Hill)
 * Elizabeth McAulay (UCLA Library)
 * Andrew Rouner (Library, Washington University)
 * Kevin Hawkins (University of Michigan)

COMMUNICATION
--
 * TEI-LIB
 * TEI-C Wiki
 * Conference calls?
 * Possible Fall meeting at DLF Fall Forum 2008 in Providence, November 2008
 * SIG Meetings at TEI Members' Meeting in November 2008

ACTION ITEMS

 * Update the Best Practices Guidelines to reflect P5 (to be assigned via working groups)
 * Ensure all levels are P5 compliant; where examples exist update them to reflect that compliance; where examples do not exist, supply them (volunteers for this task will be divided up according to the different levels)
 * Flesh out the TEI Header section of the Guidelines and update to correspond to P5 (Kevin Hawkins and Melanie Schlosser)
 * Consultation on the relationship between METS and TEI (Jenn Riley)
 * Create new wiki spaces for: (Michelle Dalmau)
 * TEI Header and Best Practices
 * Best Practices Levels
 * TEI Tite
 * Natasha to send MARC to Header mapping spreadsheet
 * Check links to existing examples in the Best Practices; update when possible (Chris Powell)
 * Identify/"massage" differences between P4 and P5 in light of the Best Practices
 * Map customizations for Level 4; Create an ODD (Syd) [Syd, can you clarify what was meant by this?]
 * Once Tite Survey is released, attendees will disseminate survey at the local and state levels (Michelle will lead)
 * Send information about the Internationalization Tag set -- ITS (Syd)
 * Contact DLF for sponsorship of conference calls or another joint meeting at the Fall Forum in Providence, RI (Matt and Michelle)
 * Create a survey to determine workflows to see how the BP levels coincide and help better define each level (Michelle)
 * Present updated version of Guidelines to Council and TEI Board for vetting and inclusion on the TEI-C web site

WORKING GROUPS
Best Practices Guidelines Working group
 * TEI Header (Kevin Hawkins and Melanie Schlosser)
 * Level 1 (??)
 * Level 2 (??)
 * Level 3 (??)
 * Level 4 (??)
 * Level 5 (??)
 * Assess issues with conformance

Priorities for the Best Practices Working Group
 * Each encoding level and the header section needs to be updated to comply with P5.
 * Examples either need to be updated or gathered for each level.
 * Create an ODD for each level? [Do we need an ODD for each level? Or one ODD for level 4 that all other prior levels could use?]
 * Note areas in workflow where METS or other standards and processes can be cited
 * Reconsider the goals and aims of each level in light of changes in workflows, TEI Tite, etc.

WELCOME/INTRODUCTIONS

 * [Round table introductions]
 * [Review of agenda]

DLF BEST PRACTICES GUIDELINES

 * DLF Best Practices Guidelines has been a work in progress since 1988
 * Guidelines will soon be under the purview of the TEI in Libraries SIG
 * Moved guidelines to the TEI wiki for revising/updating (http://www.tei-c.org/wiki/index.php/TEI_in_Libraries:_Guidelines_for_Best_Practices)
 * Need to address some issues of tag abuse that arose based on case studies of early projects like Making of America
 * The use of over  is problematic; but  is not supported by TEI Lite. TEI Lite can not be extended, but it can be modified.
 * Levels designated (1-5) are part of a process; originally meant to transition from level to level.
 * Need to provide guidance from Level 1 to METS and Level 1 to another scheme corresponding to level 2.
 * Guidance using METS since Level 1 TEI markup is bare bones structure.
 * Chris Powell has experience/documentation in converting Level 1 and Level 2 TEI markup to METS
 * IU has transformed Level 3/4 markup to METS; Indiana Magazine of History serves as an example
 * Level 1 should have 2 options: If organization plans to enhance TEI, then use TEI as the example; if organization as no plans to enhance the TEI; scrap the TEI for METS
 * Assumption that newcomers will start at Level 1 and eventually "grow" to Level 4 encoder is probably not accurate. Re-think levels as a "stepping stone" to the next level.  Consider cost-effectiveness of each level, especially if additional in-house work needs to occur.
 * TEI to METS should permeate (e.g., examples) all the levels [We can talk about this, but we should probably get the Guidelines up to P5 compliance first]
 * Focus on P5 when updating the guidelines. The differences between P4 and P5 are subtle up through level 3 or possibly 4.
 * Flesh out the TEI Header
 * Need to revisit and update encoding examples for each level; find examples for levels when and if necessary
 * Revisit the Levels conceptually, include Tite

TEI HEADER
(See TEI Header)
 * DLF Best Practices and the Header; the Header documentation needs to be fleshed out and P5 compliant
 * General discussion of the value of using the TEI Header as authoritative source especially when alternative metadata exists (e.g. MARC records for books)
 * Develop Guidelines for a decent header and find ways to link to other metadata formats in the header as part of the Best Practices
 * Arguments for a more complete header:
 * only specialized data in our records
 * makes the document more portable
 * The header should point to authoritative or alternative metadata records; needs to be supported in the header. Recommendation to the council.
 * Any known examples of headers pointing to authoritative metadata? [None reported]
 * MONK project – something that wants to gobble up a bunch of TEI – maybe they would be uninterested if the full metadata isn't there. MONK both harvests and also receives disseminations. If it's disseminated, then we can add in full metadata.
 * OAIS model is important; archival v. dissemination
 * Need recommendation on "packaging" for dissemination; send it out with complete headers, current bibliographic info, etc.
 * Need  or some such element in the header to point to other metadata formats
 * to point out to other metadata formats
 * METS wrapper -- [NOT SURE WHY THIS IS LISTED HERE ... MAY REMOVE]
 * Kevin and Melanie to work on various headers; Jenn to consult on METS model
 * Start on TEI-Lib list for continual discussion about best practices
 * Email Natasha with wiki page link so she can attach a metadata mapping Excel spreadsheet (TEI Header and MARC)

NEXT STEPS

 * Release "final" versions of the Best Practices to the TEI Consortium web site
 * Editable version of the BP document is maintained on the TEI-C wiki
 * Write a proposal to Peter Brantley about the position of the TEI Taskforce (merging with the SIG?)
 * Considering a print-on-demand model for updated guidelines.
 * DLF to perhaps support conference call; support publication of best practices
 * Michelle to generate discussion with the Levels (main issue about Level 3); Level 3 doesn't quite fit.
 * Appropriate for born digital conversion -- NEED TO VERIFY THIS COMMENT. REFERENCE UNCLEAR.
 * Set up space on the wiki for discussion on the Levels ...
 * How do the levels fit for cost model, material ...
 * Case study of workflow -- level 1 or level 2. Do people start at level 1 anymore?  Most probably start at level 3.
 * DLF in Providence (another meeting there): Nov 12-14, 2008
 * TEI Annual Meeting in London, Nov 8th.

TEI TITE

 * The impetus for Tite was to produce a tangible membership benefit. A keyboarding the most expensive part of working with a vendor is setting up the spec. The job of Tite to be the spec. How to come up with that spec was the job of the TEI in Libraries to point out the three specs that were used: CDL, UVA, UMich. Perry's job was to review those three specs and find the commonalities. Fall of 2006 was when Perry did that work and it was submitted to the TEI and it came into the fold. Sebastian and others made sure that it was kept up with the development of the P5.
 * Tangible benefit membership to the TEI; bulk discount on keyboarding vendors
 * Create a standard spec; funnel projects following that spec. to the vendor.
 * Vendors can then be up and running much quicker following one spec
 * Tite is based on encoding guidelines developed by: Univ Virginia, Michigan and CDL (union set)
 * Tite is maintained by the TEI;
 * ODD file
 * Tite has accompanying stylesheets created by Sebastian Ratz that maps TITE to TEI P4 and P5
 * Mellon funded survey to understand outsourcing practices and TEI membership benefits
 * Perry Trolard, John Unsworth, and Dan O'Donnell are on a Mellon grant to try to bring the point of Tite into being – getting a vendor discount. So we put together a survey. The goal of the survey : would you participate? If you were to participate, what kind of page number would you put into this system in a year? * Survey aimed at small to medium size e-text projects
 * Survey is 3 weeks away from going live from the survey
 * Concern about Tite being publicly available: We don't want to put the vendors in the position of determining who is a member. Is there any concern about non-members using the TEI Tite spec and getting the same discount without joining TEI, which was part of the approach.
 * One option is to keep the TITE spec publicly available, but provide the style sheets for members
 * Mich gets contacted about encoding/OCR by the state and region all the time. They can help disseminate the survey at the state and local levels.
 * Group vouched to spread the word in their respective locations in hopes to target small to medium institutions.
 * TEI-C serves as aggregator – good for the vendor. One point of contact. From the TEI's point of view, it puts a lot work on the TEI. The only benefit for the vendor would then be that they're using the same spec. if each institution manages their own quality control process.
 * Questions about whether the TEI-C will be well equipped or would want to take on role as "physical" aggregator of content.
 * When the document comes back from the vendor, then you customize it to fit your project needs. Use case we were imagining was getting a Tite document back and then customizing once you get it back: completing the header and additional semantic markup.
 * Surveying the current TEI membership is important.
 * The Mellon grant assumes there will be adoption of Tite and that John Unsworth and Dan O'Donnell will arrive at a successful negotiation.
 * RLG established a similar agreement with vendor [Apex] for EAD markup. Merillee Proffitt could provide advice in this area.

[Side discussion about OCR and Michigan]
 * Mass OCR digitization; track dollar amount (2 cents per page in-house, a little more outsourced) (text file that correspond to a page; dirty OCR; using prime recognition)


 * Anonymizer system for interfacing between the TEI-C and members with vendors (handling quality control, answering specific vendor questions, etc.)


 * When TITE was first announced, there was very little response from the community [Some discussion about folks not recollecting when the announcement was first posted]
 * Paul Shaffner (follow up with Tite), Cassandra Williams (follow up) and Lisa McAulay were supposed to review Tite and share feedback via the TEI-LIB list.
 * Syd worries that Tite is out of synch with the latest P5 release. For example, rendition created after Tite.

[Discussion of whether Tite should be conformant. Is it important?]
 * Probably not. Because the documents won't remain in Tite. (Tite doesn't have to see the light of day). We'll use a conversion when it comes back.
 * The point is to save money -- Tite doesn't need to be conformant. Documents will not be preserved in Tite and will be converted to a conformant TEI document.
 * We need to share experiences about working with vendors.
 * Discussion of saving money with just using short elements to save keystrokes.
 * Balancing the need to meet lots of projects' needs and still be standard enough to streamline work for a vendor.
 * Tite can be a customization, not conformant TEI.


 * Should TITE reference Internationalization Tagset (W3C) should be included? Syd will send info about ITS (explain).


 * Need to have the tools available for transforming Tite.


 * We should all make sure to spread the survey request to as many as possible especially at the local/state level.