Re: Datatype and identity constraints proposal of the day (17 May)

[James Clark <jjc@jclark.com>]

> > 2.  Should it be possible for parameter values to be context dependent?
If
> > so, a parameter would be modelled as a <NCName, string, context> triple
> > instead of a <NCName, string> pair.
>
> Only if we can agree to get rid of QName. But that means we can't write
> RELAX NG grammar by RELAX NG, because it uses QName.

I don't understand your comment here.  Consider something like:

<data type="whatever:foo" xmlns:ns1="http://www.example.com">
   <param name="bar">ns1:baz</param>
</data>

The datatyping service "whatever" will get passed a local name "foo" and a
parameter "bar".  The issue is whether for the "bar" parameter the
datatyping service, in addition to the string "ns1:baz", also gets passed
the namespace declarations in-scope on the <param> element so that it could
interpret ns1:baz as a QName.

James