Re: Request for clarification on Canonical XML

Hello Joseph,

I have removed the RDF Core WG and the I18N IG because this doesn't
really concern them.

At 10:53 03/07/28 -0400, Joseph Reagle wrote:

>On Friday 25 July 2003 14:50, Martin Duerst wrote:
> > >Not sure what this could mean - you mean something
> > >like a qname attribute value?  Oh.  I suppose you
> > >could work around that by adding the prefix to the
> > >inclusives list, but I see your point.
> >
> > In a very specific application, the 'inclusives list' works well.
> > If you want to be general, it doesn't work well.
>
>I'm not confident of the scenario/requirement yet,

It wasn't intended as a requirement, just as an idea.


>but *if* Alice wanted to
>ensure that a prefix is always declared by some other person (Bob)
>corresponding to a attribute value, she could use an attribute with the
>same prefix to force its "visible utilization."

Yes, that's the idea.


>However, I'd expect the
>interface between Alice and Bob would be sufficiently close in the cases
>that I think of, that she can pass the prefixes corresponding to QNAMEs to
>Bob;

Well, I don't know all the cases and how they might work. But it may
turn out that this is the easiest way to pass the information.
Just handing over a file is much easier than handing over a file
and some other information for which there might or might not be
a defined format.


>or if Bob really cares (since he's doing the c14n/signing), he can do
>the work of scanning for QNames and adding them to his InclusiveNamespace
>Prefix list himself...

Which he still might have to do if my idea is not established as a convention.


Regards,    Martin.