[tei-council] @TEIform

[Syd Bauman]

> James and I suggested to each other today that Roma should spit out
> an XSLT script for a given ODD which performed the canonicalisation
> for instances against that ODD. How would that suit?

Well I like the idea a whole lot, but only if it's used to
canonicalize that which can be: restrictions and isomorphisms. True
extensions need human intervention and should be left alone. (By this
auto-canonicalize software, at least.)

Thus the auto-partial-canoncializer stylesheet should leave added
eleements and attributes alone. Changed datatypes will be a pain in
the keester, as it will be at least very difficult, if not
impossible, to determine whether or not the new datatype defines a
set of values that is a subset of the canonical one's.


> The I18N stuff requires this script generation anyway, probably.

I think this is its main use. And also for projects that create
<chapter>, <section>, and <subsection>, all analogous to <div>.