Lists Home |
Date Index |
From: "Matthew Gertner" <email@example.com>
> I agree with everything you wrote. In particular, I see the issue of the
> perceived W3C ownership of the space. At the same time, there is also
> widespread perception that XSD has serious flaws, especially the lack of any
> attempt to find "80/20" compromises
But most conservative documents can be happily modeled using most schema languages: including DTDs. This blocks off the "it doesn't work" justification
for having an alternative, which is the big thing that can cause people to
shift from one technology to another. If the differences in power between XML Schemas and RELAX NG are largely in areas that people don't care about, then RELAX is not compelling.
What a schema language really needs is not just elegance or humble scope, but that it provides some unique capability that opens up some particular approach for markup. For example, RELAX allowed set operations on it: Murata-san has been investigating this for years, with the idea that this can allow automated management and composition of schemas, if I understand it correctly. A little of this requirement (but not the whole hog) seems to have
filtered into XML Schemas, in that extensions are only allowed by "suffixing" the content model.
I am not sure whether RELAX NG has kept up with this aspect of RELAX. It would be really interesting to know (or to know which features of RELAX NG will cause problems for this.) But I do know of a commercial product being built using RELAX NG, and I expect we will see more over time.