[sc34wg3] The TMCL schema topic
Lars Marius Garshol
larsga at garshol.priv.no
Fri Mar 27 08:02:42 EDT 2009
* Lars Heuer
> May I ask why the schema topic was dropped? Sorry, I missed the
> discussion, any hints appreciated. :)
We agreed in Prague that there was no good reason why, and it's now
> To the schema topic: *If* we have a schema topic, I wonder if we
> couldn't attach the constraints like "tmcl:topictype-constraint", and
> "tmcl:associationtype-constraint" to that topic.
> If I understood TMCL correctly, these kind of constraints enable /
> disable some behaviour of the TMCL processor.
We killed the representation of these constraints as topics in Prague.
Instead, what we did is to just define them as constraints that always
apply. The idea is that to be conformant a validator has to be able to
validate *all* constraints defined in the spec, including these.
However, it's perfectly OK for validators to either turn them off by
default or allow users to turn them off somehow. How that's done the
spec doesn't say anything about. The requirement is that it must be
possible somehow to validate all constraints.
> If my understanding (described above) is correct, I think it would be
> much more nicely if we can attach this kind of information to the
> schema topic (i.e. via associations where the schema topic plays a
> role) instead of this keeping this:
> ttc isa topictype-constraint.
> we could enable it via templates:
This issue no longer applies. :-)
More information about the sc34wg3