Re: Error in xmldsig REC

On Friday 31 May 2002 09:32 pm, Martin Duerst wrote:
> At 15:59 02/05/31 -0400, Joseph Reagle wrote:
> >While looking at xenc's use of the MimeType and Encoding attributes I
> >noticed that in the text we say Encoding="base64" but the schema says
> ><attribute name='Encoding' type='anyURI' use="optional"/>.
> Your mail is a bit out of context. Are you talking about XML Sig,
> or XML Enc? And which element(s)?

Both. I noticed the error in xenc, wondered where it came from and found it 
in xmldsig. So one way to answer my question is, what do we want for xenc? 
Do we want:

  <EncryptedData MimeType="text/xml" 
    Encoding="base64"/>
or
  <EncryptedData MimeType="text/xml" 
    Encoding="http://www.w3.org/2000/09/xmldsig#base64"/>

(This confusion is the result of my occasional attempt use actual 
identifiers instead of non-qualified strings and get IETF/IANA to assign 
such identifiers... Didn't manage to do this in xmldsig or xenc, maybe next 
time around!). On the xmldsig front, I think it'll be easier to change the 
example in the text than the schema, so that's the likely path I will 
pursue.

> it is very worthwhile to make sure that in the text, every
> single instance of the word 'encoding' is qualified
> (e.g. 'character encoding', 'transfer encoding',...), and
> that the differences are pointed out clearly.

Yes, my nativity that the task of representing and sending characters 
would be straight-forward seems inexhaustible. <smile/> I thought some 
ambiguity for the Encoding attribute wouldn't be too harmful as it's a 
informational piece of information used at the applications discretion. 
However, the syntax difference between the schema and the example in the 
spec do need to be remedied.

Received on Monday, 3 June 2002 12:39:35 UTC