xml:id RE: [sc34wg3] Compact syntax requirement question

Bernard Vatant sc34wg3@isotopicmaps.org
Thu, 21 Jul 2005 09:56:02 +0200

> > Content is the author's responsibility. If I want my XML file to
> > make sense, I'm responsible for making sense, and there's no schema
> > language on earth that can enforce that for the kinds of information
> > with which I work.

> So it is all well for you to expose 'authors' to the enormous overhead
> of creating XML without actually helping them with the REALLY
> IMPORTANT ISSUE, namely whether what they create makes sense in terms
> of the model they are working against?

Sorry, seems that I've lost the thread, but how will CTM be different for that matter? I
understand that it will need specific parsers. So instead of the 'enormous overhead of
creating XML', which is, you have to admit, sort of biased appreciation, you will have the
'slight overhead' to get and learn new tools, new syntax, new ways of validation, and so
on, and your stuff will be only usable against the TMDM model and TM applications.

One thing that has not be mentioned here is that a piece of XML simply valid against the
XTM DTD is maybe not valid against TMDM, but it's valid XML altogether, so for people who
don't care too much about being conformant to TMDM, it's a good piece of XML that can be
crunched and parsed with all sorts of available tools. The same with an RDF file, or an
OWL ontology. Of course there is a data model behind, but before all it's structured
information, and no data model or schema will prevent people to parse it, transform it,
trim it and extract at will bits and pieces that make sense *for their application*, and
which are maybe completely non-conformant to the original data model. Who cares? The same
way, when I have to import data legacy in Mondeca ITM, whatever the original data model, I
trim, cut, transform, process the data by all means that seeem to make sense in the target
application (remind you Mondeca's motto is "Making sense of Content"). The final judge
about the fact that it makes sense or not is not any schema language or data model, but
the end user.

Do you think all RDF/RDFS/OWL hackers around mind if their stuff is valid against the RDF
abstract syntax and semantics, which 99% of them (I for one) are completely unable to read
and understand (supposing they know about it)? Do you really think TM hackers are better
than those guys?