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

[Murray Altheim <m.altheim@open.ac.uk>]

Bandholtz, Thomas wrote:
> Couldn't the OASIS TM community *recommend* an XSD version?
> There are many use cases when a schema is necessary today - do we like 
> this or not.
> We should not wait until all the "rest of the standard settles out".

I think you pretty much make the case against doing this: the very
reason not to standardize an XSD version of XTM is that XSD is still
not entirely stable, especially in terms of implementations. That
XSD may be useful is not an issue when we're talking about
standardization. How can you use an XSD version of XTM if it performs
differently on different tools? That's the opposite of useful, it's
dangerous.

> I think Max Voskob has posted a valuable draft yesterday. It looks quite 
> reasonable as a schema.

By what metric do you judge it being "reasonable"?

> But I am not the XTM expert. Did anyone test it with a given XTM?

What is needed is not a single XTM document but a test suite. This
is something that OASIS has done in the past, and it can and should
be done for XSD and RELAX-NG versions to be sure that they truly
conform across multiple implementations and across the test suite.
A schema is definitively useless if it can't be relied upon to
properly constrain documents.

I'm suspicious of the use of schemas to constrain topic map
documents to begin with, as I've said before. The XTM 1.0 DTD
defines the XTM markup language and can be used to validate that
a given XTM document's markup conforms to the DTD, quite
different than what I'm hearing XSD is to be used for, which is
constraining the contents of instances -- such use is misleading.
Because of the various merging behaviours of topic maps, their
serialized instances can't be treated as static XML documents,
such that XSD is less than useful in determining, say, whether a
given topic or association has the correct form or contents if
this is taken as "topic map processing."

Proper constraint/validation of topic map structures needs to
occur at a higher level than markup, using a topic map templating
language, something yet to be finished (but in the works). This
is not to say that an XSD or RELAX-NG schema is useless, but not
as a constraint language on topic maps except as they are parsed
into a topic map engine for further processing.

Murray

......................................................................
Murray Altheim                  <http://kmi.open.ac.uk/people/murray/>
Knowledge Media Institute
The Open University, Milton Keynes, Bucks, MK7 6AA, UK

     "In Las Vegas Mr Gates also demonstrated a prototype
      fridge magnet which can be programmed to receive traffic
      reports, sports results and advertisements from local
      restaurants using the same FM signal as the wristwatch."
                                  -- The Guardian, 10 Jan 2003.