taken by Peter Ogden

Marty Sachs (IBM)

Heiko Ludwig (IBM)

Jean Zheng (Vitria)

Dale Moberg (Cyclone Commerce)

Peter Ogden (Cyclone Commerce)

Himagiri Mukkamala (Sybase)

Bob Haugen

Nikola Stojanovic

Marty said that there will not be a conference call on 3/27/02.

Marty stated that he has not heard from Kartha and thus does not know the status of his work on the NDD.

Work on the BPSS instance document for the negotiation process has not begun yet. Hima asked what pattern BPSS instance doc will be based on. Marty hoped that it would be based on the Negotiation Pattern, augmented by discussions of this group.

Jean sent out update to her Message Definitions document this week; she has not received any subsequent comments. Dale said he has some comments and will post them to the list.

Heiko posted a new version of glossary. Hima asked to add item for "negotiation BPSS". Heiko agreed. Marty asked for agreement that the Glossary should be added to the requirements document, and the requirements document considered "done" – everyone concurred.

Bob Haugen said that ebTWG is working on a new draft of negotiation patterns. Marty asked him to work with this team (in particular, Bryan Hayes and Pallavi Malu, who also represent the BPSS team). Bob said that in ebTWG there is a Collaboration Patterns group, which includes all of the original authors of the Negotiation Patterns paper, and they will work with negotiation team. The ebTWG Collaboration Patterns group is really focused on "state alignment"; there are many considerations there that our negotiation process may safely ignore.

Dale recounted some of the recent discussion with Bob Haugen and others on the listserver related to the negotiation process. He stated that this group had been reviewing Haugen et. al.’s initial Negotiation Patterns document, and pondering whether it fit into a simple request/responses paradigm. Dale stated that we’re interested in understanding negotiation so that it can be automated. He raised Marty’s and Heiko’s concerns about race conditions – i.e., how to hand off control of the counterproposal process. Dale then stated that Bob’s process contained three possible responses (accept/reject/counter-pending) and that understanding the counter-pending state had helped Marty and him to realize that the race condition had already been considered.

Dales went on to point out his main criteria for the negotiation process:

1. It needs to lead rapidly to convergence (agreement/disagreement)

2. It needs to employ deterministic algorithms and avoid infinite loops

Bob said that process should recognize when a particular negotiation "state" has been reached previously and kick the process over to a human. Dale stated that algorithms/facilities that can do this have high computational overhead and would rather avoid it – process should work on what’s in front of it.

Dale went on to suggest that when a proposal is rejected, the rejecting party does not want to see that same proposal again. The mechanism of preventing repeat proposals (and infinite looping) needs to be considered. Also, need to consider what happens when proposal is acceptable, but receiver thinks he can do better. Dale illustrated with the following example:

Dale will document this in more detail and post it to the list.

Dale wants to make sure that we take into account cases like the above (raised tacitly by Duane Nickull last year in Vancouver) and come up with a solution that strives to arrive a the "best" possible position for both parties.

Dale suggested that we need some scheme for "rating" the different acceptable negotiation results so that the process can evaluate what’s best. Bob said that this "preference rating" approach would blaze new trails (that is, it is not something that’s being investigated by the ebTWG).

There was a brief discussion of "taking turns" approach to negotiation versus "nested proposals". Bob said that the original Negotiation Patterns team had not considered the nested approach because it couldn’t be done using apparatus available in BPSS at the time.