On 3/13/06, Scott Cantor <> wrote:
> Yes, I could not find anything in the conformance doc also.
It's missing. The intent was that type 0x04 is MTI.
> So, I could define
> Artifact = [Typecode MyFooBar(nn)]
> Where TypeCode := 0x0100 (e..g)
No. All SAML 2.0 artifacts have an index field. The part after the index is
open.
Furthermore, there isn't a formal registry of artifact type codes, which
means there is no guarantee you won't step on somebody else.
PKB: Agree. I think we should provide a framework for provider defined types. That's for future discussion.
Finally, it's just generally a bad thing IMHO. What does your artifact do
that requires its own type?
PKB: In general I agree. The situation is little different for me - it
is very limited circle with lots of security requirements (Can't go
beyond that). One problem is you have to keep a map of artifact and to
which user it was issued. This requires a database or other distributed
mechanism when you have colos worldwide. I am exploring right now ...
Thanks,
/Prasanta
-- Scott
