User:Kshawkin/My Council obligations/Plan for TEI Tite

From TEIWiki
Jump to navigation Jump to search

Background

TEI Tite was created as a result of a grant from the Mellon Foundation and given to the TEI-C to host in its repository along with other customizations. Since outside support for its maintenance has not been secured, maintenance fell to the TEI Technical Council. The Technical Council decided in September 2012 to maintain Tite, as it has long done for TEI Lite.

Those involved with setting up AccessTEI agreed that Apex CoVantage would need a stable version of Tite to train staff on, so Apex forked the Tite spec to make some last-minute modifications to Tite before the launch of AccessTEI. It was always the intention to bring these revisions back into the canonical Tite source managed by the TEI-C, and in September 2010 the chair of the Technical Council appointed a task force to reconcile the two versions in use. Greg Suprock at Apex confirmed on 2012-07-20 that Apex is willing to begin using a revised version provided by the TEI-C. Kevin Hawkins gathered the DTD used by Apex (hacked from a Tite DTD generated on 2010-03-02), which Virginia McClure sent to Dan O'Donnell in April 2010, and a summary of the changes, as reviewed by and approved by representatives of the TEI, to compare against the canonical one maintained by the TEI-C to see what revisions needed to be made in the canonical version to reconcile the two.

In later correspondence with Kevin, Greg sent Kevin the DTD that he said Apex was using as of 2012-04-17, which turned out to be derived from the version of Tite released with P5 1.7.0 (2010-07-06). However, it was missing <handShift/> and included <biblScope>, etc. even though these contradicted changes made to the Apex prose documentation for AccessTEI at some point after February 2010, when Apex summarized various suggested revisions to Tite to Perry Trolard and maybe others. Greg responded in an email to Kevin on 2012-08-08 that Apex was still using an older version of the DTD though had begun to review of a newer DTD (presumably the one that he had sent on 2012-04-17).

Kevin Hawkins's intention

Kevin intended to first:

a) verify that all revisions by Apex had been made in the canonical version

and then do the following (in any order):

b) Implement other bugs and feature requests related to Tite

c) Rewrite Tite as an new-style inclusion ODD, as was done for Lite at revision 10700 (not at revision 8777 as the commit comments would have you believe).

and then finally:

d) Notify Apex that the TEI-C has caught up on updating Tite and ask them to begin using the latest version.

What has happened as of October 2014

a) Kevin implemented some tickets relating to Tite based on the notes from just before the launch of AccessTEI:

Furthermore, Sebastian fixed a bug in the Tite ODD that included titlePage and titlePart from the wrong module at revision 9883.

In June 2012 Sebastian ran a comparison of the 2012-04-17 file against P5 1.7.0. I now realize that what he reported as (b) in his message is simply the result of revision 9883.

Then in July 2012(?) Kevin diff'd the Apex DTD generated on 2010-03-02 against Tite from P5 1.6.0 (the closest version), writing annotations on a printout.

b) A few of these tickets have have already been implemented:

In addition, some other have been made to Tite as a result of changes to P5:

Next steps

a) Compare Apex's DTD generated on 2010-03-02 (and not the later one they sent) against Tite from P5 1.6.0 (the closest version) for differences that can't be accounted for simply by changes to P5 (and would indicate other revisions from Apex that we should consider for canonical Tite). One could use NekoDTD or DtdAnalyzer if the diff function in <oXygen/> isn't sufficient. Kevin began this work (the "printout" mentioned above) and will pick it up; simultaneously, Syd will use a DTD comparison tool or two to check the files in an automated fashion.

Kevin has reviewed his printout more carefully. It seems that all changes have been accounted for except that Apex's version includes @from and @to on time despite our notes from them saying that it was note needed (see bug 3136934 (old system) / 247 (new system)). Kevin will inquire with Apex on whether they really meant to keep these in. (Kshawkin 05:08, 14 November 2014 (CET)) Kevin just emailed Apex to inquire. (Kshawkin 19:41, 23 November 2014 (CET))
Ravi at Apex emailed Kevin on 2014-12-03 to say that they do not need @from and @to on time. So the canonical version of Tite should now include everything that Apex needs. (Kshawkin 01:34, 4 December 2014 (CET))

b) In any order:

-- 1) See remaining issues.

Kevin has fixed corrigible errors but then emailed the Council for guidance on whether to proceed unilaterally. (Kshawkin 00:55, 7 December 2014 (CET)) Council members suggested Kevin implement when he feels certain and seek guidance from Council when not.

-- 2) Rewrite Tite as an new-style inclusion ODD, as was done for Lite at revision 10700 (not at revision 8777 as the commit comments would have you believe).

Or was this already done at revision 8764 (aside from all the whitespace changes)? (Kshawkin 22:24, 19 November 2014 (CET))
Martin Holmes emailed tei-council to say that Tite is now an inclusion-style ODD (the new style). (Kshawkin 20:40, 7 December 2014 (CET))

c) Notify Apex that the TEI-C has caught up on updating Tite and ask them to begin using the latest version.

Ravi Thota at Apex wrote to Kevin on 2015-01-04 to say they would look at https://sourceforge.net/p/tei/code/HEAD/tree/trunk/P5/Exemplars/tei_tite.odd but wait for the next P5 release before changing the schema they use. (Kshawkin 20:55, 8 January 2015 (CET))