[tei-council] @TEIform

[James Cummings]

>Sebastian Rahtz <sebastian.rahtz at oucs.ox.ac.uk>:
>>We now know how to do this better by reference back to the ODD source,
>>which also covers attribute names.
>>Hands up all those who would be willing to quietly drop this dinosaur
>>artefact?


First, let me say that my gut instinct is to drop it.  Second, by way 
of playing Devil's Advocate, let me ask why it was there in the first 
place?

My assumption has always been that in document instances which 
conformed to an extended DTD where TEI elements had been renamed, a 
processed version of the document might exist out of context from its 
DTD.  (Since it is the processing of the document in conjunction with 
the DTD which provides the attribute.)  So once a document instance is 
moved somewhere else from it's DTD, that I have a <chapter> element 
which is really just a type of <div> becomes problematised because I 
have not documented that it is just a <div type="chapter">.  However, 
if I have a @TEIform, then the processed/expanded document, even in 
absence of its DTD will have @TEIform="div", thus providing some 
perhaps crucial information to someone happening upon the document 
years later.

What I'm wondering about the reference-back-to-ODD-source method is 
that it depends upon the existence of another file (much like, you can 
argue, the unprocessed file does to its DTD).  I'm more than happy to 
drop @TEIform, because it has never been any actual use to me, 
however, I am wary about just zapping it without being sure that we're 
providing as reliable a mechanism since Lou and others must surely 
have thought it a good idea at some point.  Well, that is the best I 
can do as a Devil's advocate I think.

So explain to me again how the reference back to the ODD source works? 
  I'm of course in favour of that as a mechanism since it also covers 
attribute names.

-James