Moving a Thread to the List: KeyInfo Extensibility

Brian and I have been working on clarifying the use of the <ANY> within the 
KeyInfo children (PGPData, SPKIData, etc.) for extensibility purposes. At 
first, I thought it was an editorial issue of trying to harmonize the 
inconsistent use of ANY in the different KeyTypes. Part of these 
inconsistencies arose from a disconnect between myself and Brian in that 
Brian intended the ANY to be used to extend elements from our namespace. I 
expected ANY to be only be present to permit our types to be replaced. Both 
of these view points are reflected in different KeyTypes and I'm agreeable 
to ANY being used in KeyInfo for replacement, and in the KeyTypes as 
complements (but not replacements). Consequently, we've been trying to 
identify schema/DTD constructs that prohibit empty content:

   [a] <PGPData></PGPData> <!-- not a biggie, but silly -->

or prohibit content that is only from an external namespace (I think it 
should be under KeyInfo then):

   [b] <PGPData><foo:MyPGPData>bar</foo:MyPGPData></PGPData>

However, at this point, Brian asked with good cause why not permit [b]? We 
don't yet agree why it should or should not, so we're bouncing it back to 
the list for discussion. You can see the areas of the spec affected by this 
issue in [1], underlined in red. I'll bounce our last two messages to the 
list as well for wider comment/review.

I'll ask Don to bring this issue to a close. <smile>

[1] http://www.w3.org/Signature/Drafts/xmldsig-core/Overview.html#sec-KeyInfo

__
Joseph Reagle Jr.
W3C Policy Analyst                mailto:reagle@w3.org
IETF/W3C XML-Signature Co-Chair   http://www.w3.org/People/Reagle/

Received on Tuesday, 16 January 2001 18:24:17 UTC