[topicmaps-comment] RE: [topicmapmail] XML Schema for XTM DTD

["Bandholtz, Thomas" <TBandholtz@koeln.sema.slb.com>]

Title: RE: [topicmapmail] XML Schema for XTM DTD

[Jim Mason:]
> OASIS has no TM committees that deal with the interchange
> format of TMs or
> the definition of what it means to be a TM. That's all in ISO/IEC
> JTC1/SC34's territory.

Right. But obviously many of us need a schema and it is no good idea to have everybody make his own.
I think we do not need a TC for this. We can do it informally. First let's have a verified draft, and then think about what to do with it.

> But if any current software does
> validate its input, it's going to validate against a DTD
> since that's how
> the interchange format is officially defined.

So we cannot be used in standard web services, which is the base of UDDI, ebXML, and more.
I am sure that's not what you want.

> At this point, the only test of a schema would have to be
> whether something
> that it thinks is valid is also recognized as valid by the DTD.

Exactly. Shouldn't be too difficult. And the other way round: validate existing XTM samples with Max' schema and see what happens.

 
> But remember that having a valid XML file is not the same
> thing as having a
> meaningful TM. I can generate all sorts of things that XMetaL
> tells me are
> valid XTM files but which mean absolutely nothing to a TM
> browser (as I know
> from copious idiotic experience).

sure. Schema is a little more powerful than DTD in this field, but not powerful enough. OWL would help some steps further. Finally, it's not so much a question of the formal language. If one's subjects are "meaningless", then the whole topic map will be too, and no engine will be able to detect it :-)

Well I want some tests of Max' schema. What would be a good XTM sample for this? Max, how did you test it yourself?

Thomas Bandholtz