>I would just limit the description of such a terminology format
> to a subset that represent the main features needed in a terminology
> databank in the kind of applications we have to deal with in the TEI
> community and make explicit how the new ODD extension mechanism
> can help people expand this core.
This (ISO 16642 conformance plus borrowing of a useful subset of TBX plus
pointers about how to get full-blown TBX into a schema) strikes me as an
eminently sensible and useful strategy.
For better or for worse, most terminologists (along with a goodly proportion
of corpus linguistics folk) left TEI firmly to one side when devising their
XML vocabularies. That was a pity from the perspective of the TEI's wish to
be as comprehensive as possible. Tracing back through the archives of
relevant discussions, a lot of out-of-hand dismissal of TEI by
terminologists and linguists showed a lack of understanding of TEI
extensibility in general and of the very powerful possibilities of feature
sets in particular. Maybe the TEI presentation of these things, especially
feature sets, must take some of the blame.
But that's water under the bridge. Terminologists will not now come back to
TEI. But it would be very good indeed if thanks to the increased flexibility
offered by P5, TEI could at least partially embrace TBX and the like and
hook into ISO 16642.
That said, I do think it will be important not to allow term-bank based
considerations to have an excessive influence on encoding recommendations
for what the TEI has historically called "print" dictionaries. From time to
time people tell me with considerable confidence that the current P4 Chap 12
is too X (where X = "simple" or "rigid" or "lax" or "inconsistent" or
whatever) to encode a scholarly dictionary of any complexity. To which I can
only reply that encoding the Anglo-Norman Dictionary, apparently to the
entire satisfaction of its scholarly user constituency, turned up no
encoding problems that were not capable of solution by very straightforward
application of the extension mechanisms.
I'd be the first to agree that neither Ch 12 nor Ch13 would be an adequate
basis for encoding a full lexical database; and I am also finding that the
task of simultaneously planning and encoding another bilingual dictionary
that has to do justice to a language very different in structure and
morphology from the European models on which Ch 12 draws does indeed bring
out some quite severe limitations in the range of things that Ch 12 allows
for. But the present Ch 12 recommendations have done sterling service to at
least one important project (and maybe several more), so, like much else in
P4, they ain't broke.
Which doesn't mean that they should be off-bounds for fixing, just that I
hope their present merits won't go unappreciated when their successor is
created. Here and everywhere else, the designers of P5 have a formidable
job, like those revolutionaries whom Schiller compared to a craftsman
charged with the task of repairing the town clock without stopping it. But
since Lou and Sebastian and the other Wizards of ODD possess the arcane
knowledge of how to do just that, we can have every confidence in the