Jeff and Mike,
I have a couple of questions on the consensus items:
"cid based start parameter in the MIME Content-Type header" does not sound
at all legal according to the MIME specifications. The start MIME
parameter accepts a content identifier as its value (the part of a cid
URI after the colon), not a URI.
"Payloads are not canonicalized during the digital signing process" seems
to head away from interoperability. If any intermediate makes a relatively
small change (say, modifying the character set) to a payload, the recipient
will be unable to validate the signature. Canonicalization serves
a number of useful purposes such as reducing the impact of inconsequential
changes. (And, in most cases, XML comments are not consequential.)
I much appreciated the sequence diagrams. I'm reminded of one of
the issues from the list because the diagrams once again point out an inability
to request the business response and MSH acknowledgement in the same (asynchronous)
message. You can't create a derivative of (3b) putting the two "responses"
together without combining them with a synchronous reply.
Your thoughts?
thanx,
doug
Jeff Turpin wrote:
-----Original Message-----
From: Mike Dillon - Drummond Group [mailto:]
Sent: Wednesday, October 09, 2002 12:53 PM
To: Jeff Turpin
Subject: FW: Informal input to ebMS
Hi Jeff,
I just sent this yesterday. I will also send
more formal feedback at the end of the tests....
mike
817 875 0794
-----Original Message-----
From: Mike Dillon - Drummond Group [mailto:]
Sent: Tuesday, October 08, 2002 4:59 PM
To:
Subject: Informal input to ebMS
Hi Ian,
Attached is some input based on the current/past DGI ebMS interops.
Also, attached are some examples of sequence diagrams we have
been
building, we would be willing to expand on these for inclusion
in the spec or in some type of implementation guide...
Can you forward me the concall number for the meetings Thursday,
I'd like to dial in if that is ok with you.
Thanks !
Have fun in Miami !
Mike Dillon
817 875 0794
