Talk:Best Practices for TEI in Libraries

The following are revisions to make to the BP before making an official "release". There is a separate list of Future changes to Best Practices for TEI in Libraries.

= Test ODDs and schemas derived from them =

Test Syd's ODDs and schemas derived from them: http://bauman.zapto.org/~syd/temp/BestPractices/. Just go to that URL, download the .rng files, and create a new XML document based on the schema. So if it allows you to insert all the elements you expect to be able to insert. Syd has been asked to make the following changes:


 * in header ODD, allow only a structured &lt;publicationStmt&gt;
 * lib1.rng: &lt;oXygen/&gt; says "Errors encountered: Probably no start pattern found".
 * The only allowed child of front, body, or back *at any level* should be a div.
 * note should not be allowed at in Level 1 or Level 2
 * ab should be the only child allowed of any div (in both Level 1 and Level 2). This element seems to be missing from the schema.
 * floatingText is missing in Level 3 or Level 4 schemas.

Possible additional tweaks to the ODDs based on email threads. Need to determine whether to move to Future changes to Best Practices for TEI in Libraries:


 * Use of attribute(s) on sourceDesc/biblStruct/monogr/imprint/date is now required, and with attribute values mapping to the Dates fixed fields (per email sent by Kevin on 1/16/2010)
 * normalizing AACR2 dates for machine processing, to be shared with Thutmose project (per email sent by Kevin on 1/16/2010)

=Pending Review: Use of any P5 attributes =


 * [Prose changed] to reflect recommended/required attributes in the body of a text
 * This text may change again after the list has been generated (it's purposefully vague)
 * May add an appendix of all attributes identified
 * Lisa is compiling a list of frequenly used attributes
 * The group will need to review the list for completeness
 * Syd will constrain the ODDs accordingly

=Resolved: Publication Statement=

As of 5/26/2010, allow structured and unstructured publication statements, per Kevin: "In cataloging, I believe you don't state publication information if something is unpublished rather than saying something like "unpublished", but I don't have AACR2 with me at the moment to verify.  However, we've been trying to give people the option of creating TEI headers conforming to the BP by hand (not from MARC source) in case they want to, and I agree that the TEI way to do this is with a &lt;p&gt; element.  So I now favor allowing either the structured or unstructured publicationStmt in the ODD but saying in the prose that people they should only use the unstructured one for a statement such as your example.  Does this sound okay to others?"
 * Irrelevant since we are talking about fileDesc publicationStmt; stick with structred statement

= Resolved: Direction of pointing between note references and notes themselves =

Decide whether to change back to having &lt;ref&gt; point to &lt;note&gt; instead of &lt;note&gt; point to &lt;ref&gt;, as Syd recommended. See this ticket:

https://sourceforge.net/tracker/?func=detail&aid=2796148&group_id=106328&atid=644062

and this change to the Guidelines:

http://tei.svn.sourceforge.net/viewvc/tei/trunk/P5/Source/Guidelines/en/CO-CoreElements.xml?r1=6937&r2=6936&pathrev=6937

or, for the full story, see Kevin's email from Nov. 6 and previous quoted messages.


 * Kevin updated this in the Bp: Best_Practices_for_TEI_in_Libraries

= meeting element =

Decide whether to include &lt;meeting&gt; in sourceDesc/biblStruct/monogr/ and/or in titleStmt. (Per a change on 2010-01-15 in SourceForge, meeting is now allowed in titleStmt.) As Kevin discussed in an email sent on Oct. 12, the name of a meeting is usually included in a MARC record, but it's not distinguished from an author or editor in the same way TEI divides up the world. The essential question is: if you digitize a volume of conference proceedings, is the name of the meeting, as opposed to the title of the volume, really important enough to warrant inclusion in the TEI header? If so, we need to wrestle with the questions Kevin brought up on Oct. 12.

= appInfo and application =

Decide whether to include &lt;appInfo&gt; and &lt;application&gt; in our header recommendations. In email discussions, Syd saw them as useful, but Lisa didn't think we need them.


 * There is at least one proposal forthcoming for further work on defining the scope and usage of these elements, which have not yet reached the degree of stability desirable for inclusion in a BP document, imho LouBurnard

= Add history of version 3 to Appendix A =

Say that the text was written between April 2008 and ___ 2010 (the release date).

= Add Tite as Level 3.5 =

Dependent on ongoing Tite revisions; need confirmation from Dan O'Donnell/Perry Trolard

This was strongly recommended by Daniel Pitti in Ann Arbor because he felt certain that administrators and funders would be confused about the difference between TEI Tite and the Best Practices ("don't the libraries already have a TEI customization?"); in fact, Kevin has known this same confusion to arise among TEI Council members. While we have a section of the BP discussion its relationship to Tite, by having a Level 3.5, we can be more explicit about mapping between the two.

Mapping clarification from Kevin: Instead of actually mapping elements, Daniel wanted us to simply proclaim use of Tite as one of a number of appropriate encoding levels for libraries.

Naturally we will not be able to describe Tite the way we do other levels -- by simply saying "all the elements in the previous levels, plus the following". Tite uses different element names of all sorts. There's no point in having Syd make an ODD for Tite since one already exists. So what Kevin envisions here is a sort of "sidebar" about Tite, inserted between Levels 3 and 4 that discusses Tite in a bit more detail than we currently have in the beginning of the BP, with particular discussion of mapping between the two.

We recently had some discussion about the merits of this, so maybe we won't do it in the end. But if we do, we'll need a draft of this new sidebar. Two paragraphs are already written for you (the brief discussion of the relationship between Tite and the BP), and you can pull more information from Tite's discussion of an earlier version of the Best Practices.

Would someone be willing to write a first draft of all of this? Two paragraphs are already written for you, and you can pull more information from Tite's discussion of an earlier version of the Best Practices.


 * Can we just use what's written here (rather than link to it) and modify accordingly fpr our level 3.5: http://www.tei-c.org/release/doc/tei-p5-exemplars/html/tei_tite.doc.html#tei-in-lib-bpg. Didn't Kevin write this anyway?  If not, whose permission do we need?


 * I'm pretty sure Perry Trolard wrote this section. Shouldn't be a problem to use it.  However, we should carefully check all the assertions since things have likely changed since Tite was last revised.  Another round of Tite revisions is supposed to be forthcoming, so perhaps wait on this.

Pending Review: Revision to Tite/BP Intro Prose
To replace what's written here:

The TEI Tite customization of the TEI Guidelines was developed as a vendor specification, to support outsourced encoding of the type often initiated by libraries, archives and other cultural heritage organization. The TEI in Libraries best practices are crafted to support in-house encoding that both adhere as closely as possible to common TEI practice and library standards yet still leave room for local approaches.

If a library uses TEI Tite for outsourced encoding, it should find that converting files from the TEI Tite format to a format conforming to these best practices is not difficult. Tite files may be converted to Level 3 with some loss of granularity and to Level 4 with the addition of some markup, which still amounts to minimal human intervention. The reason Level 3 does not contain as many elements as TEI Tite is to allow for use of this level, whether for mass digitization of born-digital source documents or for upgrading Level 1 or Level 2 texts, with only minimal human intervention.

These best practices are meant to complement the TEI Tite customization of the TEI Guidelines. Whereas TEI Tite is meant for vendors who need exact specifications for encoding without room for interpretation or local practice, these best practices document how a library or other large-scale encoding project might create conformant TEI documents as applied to vendor-generated or locally-created TEI documents. TEI Tite documents are not by design complete and valid TEI P5 documents due to the fact that they are missing the TEI Header and use “shorthand” tagging to keep costs down. However, once Tite documents are transformed to TEI P5, the Best Practices could serve as a point of reference for developing the TEI Header and applying richer markup as reflected in Level 4 or 5 of these Best Practices.

For a comparison of the TEI Tite schema to these Best Practices, see TEI Tite's Appendix A.

= Revise section on hyphenation =

Revise the section on hyphenation per outcome of the discussion on TEI-L and perhaps also on how this is handled in the ongoing Tite revisions.


 * Kevin and Syd seem to agree to follow the main P5 Guidelines
 * Is this pending on Tite revisions or P5 Guidelines revisions (this section appeared under the Tite heading ...)?