RE: [ebxml-cppa-negot] Earlier posting by Bob Kearney

[Dale Moberg <dmoberg@cyclonecommerce.com>]

Comment on acceptance history: I can see the advantage in keeping

the entire history locally and sending the shared view of that

history back and forth (to keep state aligned, including view

of history). This means that a discrepancy in the view of

history could arise, however, so an error flag for that exception

might be worthwhile, IMO, if we go this way.

 

As a disadvantage, at the moment, and with a cold,

I only see bandwidth usage (given we don't mandate

using only the sent view of history--otherwise I still

have a minor security threat worry).  I don't feel that

for these specialized interactions--that occur for deployment

and maintenance purposes primarily--we should be too worried

about the resource utilization.

 

So, overall, Bob's suggestion on acceptance histor seems OK to me especially if

it helps some implementations get done!!

Dale Moberg

Bob Kearney and I have discussed his posting of 9/13/02 (below) and the
responses to it. The prior discussion resolved all matters except the
following.  Please comment. For "Acceptance History", a decision isneeded.

ACCEPTANCE HISTORY

The main reason that Bob suggested retaining the full history of
acceptances ("edit history") is because he believes it will make for a
simpler implementation than with each party just keeping it privately.   It
is not a state-tracking issue since each party knows the current state.
Please post views on this:

Should the acceptance history be just the result of the response to the
most recent counter offer or should it contain the entire edit history of
the negotiation dialogue? I don't believe that we can simply recommend
something because both parties to a negotiation would then have to agree on
which alternative to use.