October 10, 2016
- Kevin S. Hawkins
- James R. Griffin III
- Syd Bauman
- Elli Mylonas
- Martin Mueller
- Paul Schaffner
- Rather than stepping through the Best Practices ticket triage, directs attendees to refer to Issue 23
Relation of Best Practices to TEI Simple (Issue 23)
- Council requests that this be addressed
- Hawkins had advocated for postponing this in the past
- Came from a comment by Mueller
- Relationship and overlap between Best Practices in TEI for Libraries (BPTL) and TEI Simple
- Lou Bernard summarized the differences, but only in the headers
- Lot of good will in the community, not enough time to implement
- From end user's perspective...not interested in TEI in Libraries
- Distinction between TEI and TEI in Libraries
- If faculty member has project...help desired likely relates to metadata and header (rather than encoding itself)
- Hierarchical ordering in the BPTL approach to encoding no longer works
- Levels 1 and 2 are of largely historical interest (most OCR-based bulk-digitisation projects that might formerly have employed a quasi-TEI framework are nowadays likely not to.)
- Levels 3, 4, and 5
- Differences are real...but not prescriptive - these are recommendations
- Scope of TEI Simple or a Level 4 is preferred
- A single schema offered to the community (will help you most of the time)
- Real and ongoing effort has been focused upon documentation
- The TEI Council has been properly concerned with maintaining the guidelines
- Feels that the current guidelines in the BPTL are largely useless...too long and complicated
- For skilled programmers these are even too complicated (they won't read the guidelines)
- Need guidelines which speak to the average user
- These shouldn’t address customization
- Focus should be shifted to the relationship between faculty members (patrons) and librarians
- Clarifies that these are not guidelines for assisting users coming to the library for assistance
- Instead these guidelines are for mass digitization undertaken by the library
- Level 4 is the only level which has gained any traction
- This Level is quite close in scope to TEI Simple
- Not enough difference between mass digitization and the average needs of beginning users for TEI
- A separation of the BPTL document and documentation maintained by the Council for these two cases requires that both be maintained
- Not enough energy and time to maintain two documents
- Merging a TEI guide (similar to what is used for TEI Simple) with the BPTL guide would get further ahead
- Earliest versions of Best Practices used the TEI Lite DTD at all levels.
- Later, for Version 3.0, Bauman created ODD's for encoding Levels 1-4
- Document in the past has been used for large-scale encoding undertaken by a library
- Level 5 (fake level anyway) is used for advanced needs (faculty consulting with librarians for encoding in projects)
- No good data, but has a hunch that libraries aren't using the document for mass digitization (prefer to work with HathiTrust or other service providers)
- The document remains a resource for institutions committed to using TEI (and maintaining a uniform TEI infrastructure)
- This would be maintained as an alternative to working with providers such as HathiTrust
- Admits that the Best Practices document isn’t the best starting point for novices
- It assumes a certain knowledge of TEI and XML
- Not ideal for a first introduction to TEI.
- Asserts that TEI Simple also isn't the ideal starting point either
- Recommends instead that a novice start with TEI Lite
- Once we published version 3.0 of the BPTL, the Working Group asked the TEI Council to take over maintenance, but at the time the Council wanted to leave it to the Libraries Special Interest Group (SIG). They said that if the Libraries SIG didn’t want to maintain it, that might be an indication that there actually isn't a real user community.
- Everybody comes from a particular environment
- Northwestern University environment conforms to a high-end North American library
- No TEI competence in this library
- There are some who know about it
- No institutional interest, commitment, or infrastructure of knowledge
- Future projects are to involve a traditional patron of the library (e. g. student)
- Might be an encoding project
- Josh Honn (Digital Humanities Librarian) and a graduate student consult with the patron
- In this case, the document isn't of use
- Need a generic (not library-specific) introduction to the TEI
- Library is a service organization, hence, well suited for providing such an introduction
- Should this WG be disbanded in order to undertake this type of an approach?
- The purpose of this WG is to clean up the Best Practices for Libraries
- Objective is to ensure that the BPTL's are current - and some of the disambiguations that we are engaged in will be useful in documentation of other schemas as well.
- Not advocating for wiping out the BPTL's and simply documenting TEI Simple
- Recommends enlarging the problem and looking at documenting the usage of TEI in Libraries more broadly
- What kind of document should there be to address this ?
- Identifies two needs
- Introduction to TEI for Digital Humanities Librarians and faculty
- BPTL’s as light documentation on how one might encode materials involving the libraries (and to ensure interoperability with library systems)
- Library digitization projects are increasingly rare (thing of the past)
- Patrons have extremely specific interests (e. g. tying text to GIS)
- Is there even a generic user for libraries?
- Their instinct is that there isn't a generic user (but this would require research for confirmation)
- References the new SIG on periodical literature
- This is found within libraries
- That's exactly where different levels of encoding are useful
- Some periodical digitization projects are interested in metadata and some kind of chunking
- Describe levels and features to be encoded
- EpiDoc as an example for a successful schema and practices arising out of a user community
- Hasn't seen generic patrons either
- No request for printed books as use cases
- Anomalous materials typically requested, usually need ODD's and customization (e. g. Arabic grimoires)
- What is TEI in Libraries now?
- This is a broad question...unclear as to whether or not this can be thoroughly address right now
- The last point would require an entire morning in a conference room for this discussion
- Maybe even a new SIG
- This workgroup is under the umbrella of SIG in Libraries. Another WG could be formed (on getting users oriented in the TEI).
- When researcher comes in...doesn't likely know about the TEI
- Coming into the library looking to undertake research using text
- Librarian determines whether or not the TEI is the proper tool
- Clarifies that they are not challenging Mueller's arguments, just might not be the proper time
- Yet another independent solution could be effective
- But, the better approach might be to merge multiple solutions
- This ensures that the merged solutions can be more readily maintained
- References the journal schema as an example
- Product proliferation often continues when production unification is best
- Yet, BPTL’s WG is not suited to address this
- Merging TEI Simple and BPTL would produce quite a lengthy document
- Returns to the problem of having a (still) fairly complex document
- Users come in and do not know where to go.
- Asserts that the WG should go through and finish with the work for the BPTL’s
- Admits that we don't know who the users are
- Other parties can take on the additional project of a tutorial on the TEI in Libraries (Or, extending TEI Simple to serve this purpose)
- Recommend that work on such a tutorial become a new WG within the SIG
- Should draw upon the work in BPTL’s WG
- Should also involve Digital Humanities Librarians who actively work with patrons in libraries
- Not entirely convinced that this is a worthwhile endeavor
- Looks to push effort forward with a call for volunteers
- This will determine whether or not this will indeed be worthwhile
- Question of TEI Simple
- Did TEI Simple began with library-based projects?
- Is this a sibling of other guidelines based in libraries?
- If so, might reconciling these guidelines be of use?
- The experiences of Michigan, Virginia, and other institutions are indeed the basis of both the BPTL [though more precisely TEI Tite] and TEI Simple. The resulting schemas for Level 4 and TEI Simple are not far off from each other. Can probably come to agreement between multiple schemas.
- (Headers are the exception)
- BPTL’s structures headers in responses to MARC (and other cataloging standards)
- Confirms that it would be useful to convert between schemas
- Agrees to avoid touching the header
- Separate validation scheme for headers was desirable (but found to be impossible technically)
- This required that a separate processing models be formed and TEI Lite updated
- Now: Should user use TEI Simple and TEI Lite? This is a dumb question
- For the vast majority of users, neither TEI Simple nor TEI Lite should be used
- In the end Schaffner identified the real point: there are no generic users of the TEI
- Contests this: users should stay within a higher-level platform and avoid the rhetoric of customization. It creates difficulty and cuts off the user base
- IT Directors and Librarians are reluctant to touch the TEI as a consequence
- Immediately projecting 6 months for developing a schema discourages use
- Cannot solve this larger philosophical debate
- Also, cannot address the issues of TEI Lite vs. TEI Simple: neither are under the control of this workgroup.
- BPTL's document does refer to TEI Tite
- Can add statements which distinguish from TEI Lite and TEI Simple as well
- Also, can introduce harmonization of Level 4 and TEI Simple: this would be worth doing
- (Calls for interest)
- Notes that TEI Simple is under control of Council, so if any changes are required to that schema, we would need them to agree to it.
- (Adds a comment to the GitHub Issue to clarify how TEI Lite and Simple differ from the Best Practices)
- We can create a separate GitHub Issue for discussing the reconciliation with Level 4 encoding.
- (Supports this)
- References cases with EpiDoc
- They issue requests for features to the Council all of the time
- Should add prose relating TEI Simple and TEI Lite
- In Vienna (the annual meeting for the TEI Consortium) Hugh Cayless said that there’d be a new release of P5 by the end of this year. Believes that the TEI Simple is supposed to come out with this new release as well. So it would be good to address alignment before a new release of TEI Simple is out.
- These Issues are not likely to be resolved before the release of the TEI P5 or the TEI Simple
- Plan is that Council will take on the maintenance of TEI Simple on an ongoing basis?
- In that case, we can deal with Level 4 vs. TEI Simple as we have time, without the pressure to finish before December
- TextGrid developed “Baseline Encoding”. The DTA people in Berlin took this over [actually, theirs is called “DTA 'Base Format'”], and then the schema [the latter] was adopted by CLARIN.
- Just interested in sticking with a single schema
- It’s a bit tighter than TEI Simple
- Success with the CLARIN DTD for periodicals
- Request to clarify: do you mean “schema” or “DTD”?
- Periodical WG stakeholders are likely to be focused elsewhere
- We could even decide that Level 4 will be replaced with Baseline Encoding.
- But, this doesn't address reconciling with TEI Simple
- It's also possible that Level 5 becomes "pick a schema and document it"
- Level 5 right now involves customization in some manner
- TEI is plagued with problems of offering multiple solutions in approaches to encoding, so indeed, documentation or schemas or whatever that says “of he N ways TEI recommends, you should use this one” are Good Things.
- In looking at Baseline Encoding, TEI Simple, and the Level 4 for BP's in Libraries, we might even end up with a refinement of Level 4 (e.g., Level 4a and Level 4b)
- References D. W. Winnicott and Freud on motherhood
- If a mother is good enough, any further “mothering” is unnecessary
- Similarly, the "good enough mother schema" is best for TEI
- Institutions can support this approach
- Encoding level can be recommended by the BP's for Libraries
- November 7th, 2016 would be the next meeting
- (Going to be at the Digital Library Federation 2016 Forum)
- (The attendees prefer to postpone the meeting for 1 week)
- November 14th, 2016 is the next meeting
- Clarifying documentation is necessary
- Prefers that parties ask questions or comment on these GitHub Issues (between meetings)
- (Will expedite the resolution of some of these outlying GitHub Issues)
- Reiterates over the issues for creating a separate issue for addressing libraries serving patrons who are novices regarding the TEI
- Should aim to ensure that the new TEI-C website (and practices in the TEI Community) are easily understood by the uninitiated.
- November 14th, 2016 is the next meeting