[tei-council] @TEIform

[James Cummings]

Sebastian Rahtz wrote:
> This is all true _only_ if you have taken your source document with its
> DTD, and run it through some sort
> of normalization process.  The TEIform attribute of your <chapter> is
> provided only by the DTD, not
> in the instance (does anyone use TEIform manually?). I suggest that the
> using the "normalized" form of
> document, with all implied attributes inserted, is not
> remotely common working practice.

I agree, that is why I was careful to say processed/expanded instance 
document.  I know of no one who uses TEIform manually.  And yes, I 
don't think most people use a normalized document with all those expanded.

> If you keep the DTD around, then things are OK.

Yes, though after initial creation/validation there are many who 
misplace their DTDs.

> In P5 world, you have to track two support files, the original ODD file,
> and the generated schema file; the downside is
> that you may use the schema file every day, but the ODD only once in a
> blue moon, so you may get confused.

This is what makes me wary.  While I can picture people keeping their 
schemas around (as much as I can picture them keeping their DTDs 
around), once the schema is created, I find it very likely that people 
will mistakenly misplace their ODD sources.  But I suppose we just 
have to remind people that ODD is part of their project documentation 
of how they differ from TEI.

> I find <chapter>. I read the ODD source to see if if
> "elementSpec/altltIdent[.='chapter']" exists.
> If so, I grab ../@ident, and rename my <chapter> to whatever that is.
> 
> easy, no?

Yes, perhaps not as easy as just looking at the @TEIform ... but still 
fairly straightforward if I have the ODD source.

> of course, a variant ODD processor might generate you a custom XSL
> script or the like to do the job
> straight off.

Do you have an XSL already which does this for simple syntactic sugar 
renamings?

-James