Lists Home |
Date Index |
>> Or you could have postponed this errata to W3C XML Schema 1.1 which
>> hopefully would involve a namespace change.
> It may or may not have a change but why do you _hope_ for that.
> That's saying that if XML schema has any kind of revision, then the
> name of every schema construct should change.
I think that part of the issue is that XML Schema doesn't have any
version indication aside from the namespace used for the schema. So if
XML Schema 1.1 *didn't* involve a namespace change from XML Schema 1.0
then it would be impossible to distinguish schemas written for XML
Schema 1.0 from those written for 1.1. This means that schema
processors can only support one version of XML Schema; if people
upgrade the processor without upgrading their schemas (very likely
when processors are embedded in (XSLT processors embedded in) browsers
and people view (transformed) XML pages pointing to schemas), their
validation will not work as expected.
I agree that versioning using namespaces is a bad idea in general, but
if you don't have any other mechanism, you don't have much choice.
How do you handle versioning in MathML, by the way?