Sure, there would be plenty of cases where TREX would regard two datatypes
as distinct even though they accept the same set of strings. But I think
that is bound to be the case with any definition of equality, given the
complexity of XML Schema Part 2. The best we can hope for is a
well-defined, simple, interoperable test that won't cause problems in
practice.
----- Original Message -----
From: "Kohsuke KAWAGUCHI" <>
To: "TREX ML" <>
Sent: Tuesday, May 08, 2001 9:59 AM
Subject: Re: ID/IDREF strawman #2
>
> > I don't want to do this, because I don't think it's well-defined in XML
> > Schema. With my proposal, TREX could consider two datatypes equivalent
if
> > they have the same name and same set of parameters.
>
> There lies a trap. Those two might be different types.
>
> <data type="xsd:qname">
> <param name="enumeration" value="foo:foo" />
> </data>
>
> <data type="xsd:qname">
> <param name="enumeration" value="foo:foo" />
> </data>
>
> You can't tell whether "foo:foo" is a qname unless you know something
> about xsd:qname.
>
>
> But one can certainly argue that this is a corner case.
>
> --
> Kohsuke KAWAGUCHI +1 650 786 0721
> Sun Microsystems
>
>
>
>
