RE: [dss] CMS

[Trevor Perrin <trevp@trevp.net>]

At 12:21 PM 4/21/2004 -0400, Edward Shallow wrote:
>Trevor,
>
>     We've made some good progress. Once again, thanks for your patience and
>unflappable professionalism.
>
>I have always agreed with the semantics-to-syntax argument you make. I am
>just at a loss on how to make it simple. What you are suggesting re: making
>SignatureObject optional and leaving semantics for the profile works for me.

Okay.


>However, we would have to also make InputDocuments optional also for CMS.

This is a separate issue, I think.  You have 2 distinct proposals:
  - make <SignatureObject> optional for XML-DSIG
  - make <InputDocuments> optional for CMS

We seem to have agreement on the 1st (make it optional; when it's absent, 
semantics are profile-defined).


>  I also agree that the semantics I wrote up are a mouthful, but they work.
>
>Another option if you want to try something else is ... additional core
>elements that would allow us to tighten up the semantics. A trade of
>elements for semantic complexity. Maybe an optional DocumentWithSignatures
>which would be initialized by XMLDSIG requesters only if multiple signatures
>were involved. This could be complemented with an additional SignatureObject
>sub-element something like MultiSignedCMSSignature.

I would rather not add multiple-signature-verification to the core.  It's a 
big change.  I'm not even thrilled that we're allowing profiles to do it, 
but I guess it's an acceptable compromise.


Trevor