> Paul Prescod wrote:
> But if you really do want to combine two (or more) schema languages,
> RELAX NG sticks more closely to a "pure syntax" view of XML validation
> than does XSD. So the boundary would be clearer.
From: "Linda Grimaldi" <>
> I'm not sure it is that straightforward. It's awfully handy to be able
> to use XSD types to constrain RDF values ...
>XSD is a little overzealous in that it would also allow
> you to capture limited object relations, a task better left to RDF and
> its cousins (and sisters and aunts, for you G&S fans out there).
I am pretty unclear about the utility of a dimarcation between "syntax"
and "relations", with the former being the proper subject of schema languages.
A schema language specifies constraints, and may so in a declarative
way that allows other useful processing (such as type labelling) as a
side-effect. Given that there is no a priori reason to expect any database
to have all its important structural relationships conform to a tree structure,
schema languages based on grammars alone can never be expected to
provide a comprehensive solution to validation.
Cheers
Rick Jelliffe
