[tei-council] xml-colon-thing

[Syd Bauman]

> I am now reasonably persuaded that it is possible to make good
> arguments for going to both xml:id and xml:lang without further
> ado.

Does that mean you believe that all the arguments CE and SO have put
forth up to this point have not been good? I'm not sure I would
consider the mere possibility of good arguments is sufficient, at
least in the face of any good arguments against.


> Drafts need to be produced for the relevant pieces of P5 as soon as
> possible, so that Council may review them. It is one thing to agree
> in principle to an idea as sketched out e.g. in David's summary of
> the issues, but quite another to agree to its full ramifications
> when presented in Guidelines-style prose with examples and
> justifications.

While your logic (that it's worth reading full idea fleshed out with
examples etc) is sound, the conclusion (that P5 should be altered
now) does not follow.
* Changing P5 so that it refers to xml:id= is a bit of a deal, as all
  Sebastian's style-sheets need to be changed to match.
* Changing P5 so that it uses XPointers to point to xml:id= instead
  of IDREFs is a really big deal, as all Sebastian's style-sheets
  need to be changed to match.
* I don't see that there would be justifications in the Guidelines,
  anyway, so if we want more of that we need to ask for more working
  papers sort of things.
* Changing P5 to do these things is itself a very significant change
  (e.g., dozens to hundreds of examples need to be changed; at least
  10 chapters discuss pointing), so IMHO Council should be pretty
  sure of it first, before we devote serious time and effort to it.
  That said, until Baltimore, I thought we *were* pretty sure of
  this.

I find this whole discussion a little dusty (as in "the dust never
settles" :-). A few weeks before Baltimore the discussion (among you,
me, & Sebastian) was not *should* we make this change, but *how*
we should do it.

Sebastian said he would not be ready to perform the needed changes to
the style-sheets until at least December, at which time we either have
to halt other production on P5 while these changes are made (my idea)
or use Perforce to create a separate branch of P5 source and make the
changes there, re-merging them when complete (Sebastian's idea).

I realize this is a weird role-reversal (months ago I was the one
chomping at the bit to change P5 to use XPointers, and first Lou,
then Sebastian, were saying "whoa" :-), but I really have concerns
about putting serious resources into this change unless we are quite
convinced we want it in the end product.


> I am also reasonably well persuaded that there isn't any
> immediate need for TEI P4 compatability module, and unless anyone
> on Council wishes to argue otherwise, my intention is that we do
> not do so unless and until the idea surfaces again in the TEI
> community. (But mark my words -- it will).

Good. But if you really think it will surface (I'm not 100% sure I
understand why it would -- perhaps you could recreate your chart of
data vs tools you presented in Baltimore), we should keep it in the
backs of our minds as we move forward, so as to not do anything that
would arbitrarily make it difficult to do so.