Web Services Architecture Working Group
21-Mar-2002 meeting minutes

Working Group home page · Meeting records

Action Items: pending, new

Attendance

Agenda

See agenda posted by the Chair .

Detailed minutes

This first item is a change in how minutes are taken. The scribe is to be responsible for formatting the minutes to enable to post them on the W3C web site.

Agenda review Approval of minutes face to face formal Liaison with DAML organization dive in to pick up to review goals from each of the champions request for discussion of Reliable and predicable of web services. (not covered due to time constraints)

3. Approval of minutes from last week

no objections were voiced.
Hugo: there were questions about security do people have anything to say about the minutes.
Suresh: he asked what was recorded and what was recorded seemed to differ from what people remember. No real concern.
Chris: minutes approved. They will be posted to the web page.

4. Review action items

5. Status reports:

Face to face - 33 people signed up. attending or regrets - conference bridge has been set-up Teleconference is available. we are looking to have a groups dinner Tuesday night. So early arrivers can attend the dinner. Members of the architecture group can attend the other WG's dinner on Wednesday. people need to register by April 1 in order to have enough food and other materials available at the meeting. The agenda of the F2F will be available shortly.

Future F2F - in the 2nd week of June - in France - trying to nail down hosting details. Hugo mentioned that Canon sent a note regarding that. Hugo and Chris were waiting on each other for a note. can someone pay for this meeting? the amount is about 4000 euros. This would in in Brittany near the Canon facilities. If you can cover the cost contact Hugo or Chris. There was discussion in the coordination group of a vacation mouth in August. Chris was in favor of that. We are going to take the month of August off. We will resume in September. On the timing in June: the Soap group is having a meeting the first week of June.

Editing team report:
Daniel - they met Wednesday - to publish a requirements document tomorrow. All of the things Chris mentioned will be included. This is one prior version before the face to face. A lot more text and "flesh on the bones" there will be one more to be published before the F2F. Comments need to be in by the 25th of March. Anyone with threads try to provide summaries by March 25th. This is so that these threads will be considered at the F2F.
Joe: is there any leeway on the closing day do to schedule conflicts. Tracing mail threads can be difficult.
Daniel: Champions for each thread need to get their summaries ready for submission. The sooner we can get the summaries in the better we can be prepared. Joe is the champion of #6. The editors figured out what was with the issues list. This was sent to the mailing list.
Daniel - all summaries need to be in by the 25th to have it ready for the F2F

6. DAML initiative liaison

Katia Sycara has joined the call this week. She introduced herself. She is at Carnegie Mellon. DAML+OIL for semantic annotation of the web DAML+S an ontology and components for describing invoking services. BBN, Nokia, Stanford, and Yale among others are involved. Two review documents are published. They are at www.DAML.ORG/services They are aiming to publish in a month this is an ontology - a web resource can provide a services: it has a service profile describing what it does to advertise their services with a registry. A service model provides what a service does a service grounding describes how to use it. The service profile. provider, name, QoS or rating. The process model and profile have inputs and outputs. She gave a description for how these could be used. (Thompson) How are things described, is there a semantic framework? the "holy grail" the matching is in form of keywords and semantics. with tokens and strings - if the service is well described not everything has to match. there is a DAMLS matching tool with partial matching. This is to operate on top of UDDI. UDDI does not provide any of these capabilities. This semantic matchmaker will operate with UDDI. If you have questions send them to Katia. How does this description language differ from other description languages? How does this relate functionally or hierarchically with other description languages? description logic or ?? there are a number of functional choreography ;languages fit in. ebXML -m not just a language - DAML+S is on top of RDF. If you have further questions contact Katia Sycara at katia@cs.cmu.edu

7. Continue review of Goal Summaries

DG#3

Suresh: in regarding goal 7 - was that finished at the last teleconference?
Chris: goal 7 - that was not really concluded, but it was continuing being discussed. What is needed to conclude it. It needs to be resummarized. The summary of the threads should be: a combination of what are the different view in the threads and the success factors related to that goal, then part of the out come would be a refined phraseology.
Daniel: The editors would like to have the agreed on final wording of the goal, what are the success factors and requirements - in addition to a diagram (similar to what was created for goal #1 and Goal #2) There would be a proposed wording and the editors would approved. the diagrams were posted to the working group by Daniel. A picture is better than words to see how it works. These diagrams are simple Visio diagrams. Daniel went on to say to put some boxes together - as in an org chart for requirements. Daniel would like to have comments on #1 and #2. These are stakes in the sand.

Nilo: #3 - not much e-mail on #3 - to develop a standard architecture for web services that is sufficiently extensible for future expansion- separating transport from what web services are or what they do. descriptions should separate design and run time aspects. Allowing extensibility. Make extensible to enable future business goals. This should not impede adoption of further interactions and to handle future business goals. We need to enumerate facets of W/S - what are the aspects. What are the other blocks to allow the extension and addition of web services.
Daniel: we should anticipate changes so aspects can be added to web services, things we have not yet thought of. Don't make the architecture rigid. We need separation of what from how. The architecture should not require that you can not do certain things. In addition we should not depend on the current state of the web. Situations change, business goals change, web services should be able to change to support future need. we can not know the future. Separate the abstract from concrete. The architecture should be able to be changes no explicit dependencies on the current situation. We need to create an architecture that allows for change. Leave room in your architecture for change.
Nilo: last question - how is this different from #7 - predicable evolution of architecture. technology and business goals. How would it look to combine #3 and #7? this was a proposal. #7 was intended as a well defined way of change management. not for extension of the architecture.
Chris: we need to come to an agreement

Chris: #8 - Tim Jones is not on the call - no traffic in regard to this. he asked if the architecture is complete and self-contained. is this the notation for describing the architecture. the original goal are coherent and consistent. for parts to work together - and consistent - there is a common protocol for communications. It was suggested that Daniel and Tim discuss this. Continue discussion to this on the list. mike man- "in its definition" how we notate its definition. there was no intent for specific notations. Consistency of the architecture was discussed. Coherency and consistency is seen as important.
MikeM: This may overlap with #11

Mike: #11 - the current wording was read. - basically - to work with existing web architecture heterogeneous environments. This may overlap with semantic web goals. Semantic web is in planning, we have to be in tune with the current web, but not to go against what semantic web to working toward. There is some overlap with #4 platform independence and heterogeneous environments are similar.
Abbey: Mark Baker is discussing this with current looking and future looking. What is the implication of #11 and #4 separate, we know how the current architecture works, but we are not aware of how the semantic web will work. They have some frameworks, but not an architecture. We need to concentrate on what works now, and we also need to make sure that we are compatible with the future semantic web.
MarkB: it is not as clear cut as current looking and future looking. we need to be in sync with both.
Hugo: the semantic web people argue for semantic models for what we are doing, as long as both goals are met we are safe. Making the difference between the old web and the future web we need to address both at the same time.
Daniel: we need to do them both. We need to know what is the tension between what one has and what the other does not have. The semantic web will provide additional functionality. What will be different in the future? what are the contradictions? It's OK to have two separate goals as long as we have both of them. Do we need to make them a single goal? Semantic web is still emerging. We need to be careful of what are the priorities and requirements. We may need to have give and take on what we are developing, how could these be joined? or should they be? This Could be a topic for next week-s call. Mark and Mike thought this was more of a management task. more opinions are welcome on this topic. If the semantic web is a superset they need to be backward compatible with the current web. What are supersets and intersections between web services, the current web and what is seen as the semantic web. DAMLS is an offshoot of the semantic web. Are semantic web people developing a reference architecture?
Daniel: they are not that far along it is not very well defined. How can we work with their architecture if they do not have an architecture? We do not have to do every goal at once, we may not be able to work this goal until later, since we do not know their architecture, but we need to be flexible to accommodate their future architecture. To be aligned with their architecture as well as to be compatible with what they are producing.

it is 13:28 (Pacific daylight time) we are running out of time.

Hugo: The semantic web group is paying attention to what we are doing here.
??: At the F2F we need to have a talk about the semantic web. we need to follow what they are doing.
Chris: we will have something on this. Each of the members will need to examine semantic web information to understand what they are doing and how it effects what we should or should not be doing.

Chris: urged people to review the summaries of the goals posted in the agendas each week. We need to make the best use of the list and the teleconferences.

Review of outstanding action items

ACTION: Editors: Rename D-AG0006 into AG-0006: PENDING
ACTION: DanielA: post issues list into CVS repository and post pointer to it; PENDING
ACTION: Editors: include definition of web service and make it clear that it is an interim working definition and may change over time; PENDING
ACTION: Editors: include glossary definitions; PENDING

Summary of new action items

See also the list of pending action items.