Re: Draft Minutes of 24 August 2000 Teleconference

At 21:43 8/24/2000 +0000, Donald Eastlake 3rd wrote:
 >- KeyInfo/X509Data - there were no objections to the last version
 >posted by Donald Eastlake which appears to be the consensus:

Ok, placed in Editors' copy.

 >- RetrievalMethod - there was agreement with the comments by Gregor
 >that URI should be a mandatory attribute.  Donald will change and

Ok, I tried to work this in with Merlin's correction, and created new namespaces for KeyInfo element types, however, is the Type supposed to be mandatory? (See previous message).

 >turn based on the discussions during last weeks teleconference.  An
 >important question is whether a <foo:bar#mumble> URI returns an octet
 >string or (when the target is XML) a node set or the serialization of
 >a node set.

What's the action item on answering this question?

 >- qualifying attributes - After brief discussion, the conclustions
 >were as follows: our current schema calls for attributes to not be
 >namespace qualified.  We should include a sentence indicated that
 >qualifying the attributes we define is an error.  It would be good to
 >see if schema checks will flag such problems.
 >the whole document.  [I have a note that there is an action related
 >to this for Joseph but I don't remember exactly what it is.]

My action item was to determine if a schema validator would choke on a qualified attribute if we said in fact there is an unqualified attribute (or what happens if a qualified and unqualified version are both present?). As I suspect, a qualified attribute would be schema invalid: in [1] the Id attribute on the Signature is schema valid [2]. The qualified dsig:ID attribute in [3] is schema invalid [4]. Consequently, I don't think there is any argument for us to change our unqualified attributes, it makes the XPath expressions a bit shorter too.

<Signature Id="foo" xmlns="">


<Signature dsig:Id="foo" xmlns="" xmlns:dsig="">


Joseph Reagle Jr.   
W3C Policy Analyst      
IETF/W3C XML-Signature Co-Chair

Received on Tuesday, 29 August 2000 16:48:07 UTC