[tei-council] on deprecation doc (was "Re: build failure")

[Syd Bauman]

> Indeed, in Oxford we decided that I was to add a note to
> http://www.tei-c.org/Activities/Council/Working/tcw27.xml saying that
> we would announce any deprecations to the community and that people
> should get in touch if they object. I have just added this to the
> instructions, as well as standard wording for the Schematron errors
> (using what Syd ended up implementing).

I think we're agreed that the wording I threw in there is not ideal,
although we've not concluded what the wording should be. Suggestions
so far (for an element; obviously would need to be slightly different
for other constructs):

a. The TEI-C may drop support for relationGrp element as early as
   2014-06-17

b. The TEI-C will drop support for the relationGrp element after
   2014-06-17

c. The relationGrp element will be removed from TEI source code on
   2014-06-17, and will not be available in any P5 releases after
   that date.

But more importantly (and this is aimed mostly at Kevin, I guess),
the procedure outlined in the 2nd bulleted list item under
"Deprecated" in
http://www.tei-c.org/Activities/Council/Working/tcw27.xml is simply
wrong. We don't add a Schematron rule to the spec. All we do is add
@validUntil to the spec. The build process automatically generates
the Schematron rule by looking at @validUntil.

I'm also in favor of adding valid=false to the <egXML> at the time
validUntil= is added to the <elementSpec>, rather than deleting
examples from the tagdoc. (4th bulleted list item.) I notice this is
what was done for <relationGrp>.

I'd also prefer to see stronger wording for the 5th bulleted list
item, to wit something like:
   We should mention the deprecation in the release notes both for
   the release immediately following the addition of the @validUntil,
   and the first release after the date specified on @validUntil.

I'd work on this myself, but I don't know where source to web
documents is kept, and don't have privileges to put 'em up.