RE: XML Signature 2.0 Strawman Proposal

> Namespace-specific canonicalization 
> does NOT mean every namespace must have its own canonicalization 
> algorithm, just that the canonicalization algorithm can be set on a 
> per-namespace basis.

Yes, I know, and what I said is correct -- a generic signature processor 
has to know about the namespace of everything it might see, even if it's 
only a default case that says 'nothing special.'

I am also concerned about combinations; assume a b and c are bound to 
three different URI's
        <a:foo b:foo='somevalue     '>
                <c:foo/>
        </a:foo>

Which c14n rule applies, and where?

The concept turns a signature from being about bytes, to being about 
application semantics.

        /r$

--
STSM, DataPower Chief Programmer
WebSphere DataPower SOA Appliances
http://www.ibm.com/software/integration/datapower/

Received on Thursday, 8 November 2007 17:19:38 UTC