See also: IRC log
Present: Daniel_Austin, Eric_Newcomer, Frank_McCabe, Katia_Sycara, Mike_Champion, YinLeng_Husband, PaulD, Roger_Cutler, Shishir_Garg, Sinisa, Zulah_Zulah_Eckert, DBooth, Doug_Bunting, Geoff_Arnold, Hugo_Haas, Ugo
<hugo> Agenda: http://lists.w3.org/Archives/Member/w3c-ws-arch/2003Oct/0026.html
<mitrepauld> re minutes from 2003-10-02, I sent belated regrets at http://lists.w3.org/Archives/Member/w3c-ws-arch/2003Oct/0016.html
Scribe: (See also
<Scribe> ACTION: (See also pending actions listed in http://www.w3.org/2002/ws/arch/3/09/ActionItems.htm and http://www.w3.org/2002/ws/arch/3/10/02-minutes.html ) [PENDING]
MikeC: In the interest of time, we won't go through the whole list. Any notable progress?
dbooth: Discovery updates are in progress. Probably will check in today.
... Also created initial version of editor's term usage list: http://dev.w3.org/cvsweb/%7Echeckout%7E/2002/ws/arch/wsa/terms.htm
Scribe: Intermediaries work: MikeC working on it.
... Reliable Messaging: Does Hao own this? (Mike will ping Hao later.)
<mitrepauld> david, I am interested in Discovery and look forward to reviewing the update.
Scribe: Security model: No progress yet since F2F.
Katia: Abbie and I are trying to schedule discussion on this.
Hugo: We could paste Abbie's whole text into the document and force people to comment on it.
Scribe: ACTION: Katia to email Abbie about pasting Abbie's whole text into the document as a starting point and to force people to comment on it
<ugo> text for security in SOAP:
Ugo: I also contributed something to the list.
Ugo: I got it from WS-I.
dbooth: If it's from WS-I, we could not include it without a copyright release.
Ugo: Then just use it for inspiration.
MikeC: We should use it for comparison with what we write.
Scribe: Privacy status: (Mike Mahan not on the call)
<Scribe> ACTION: MikeC to ping Mike Mahan on the status of the privacy work he did.
... Semantics, Semantic Web alignment: Any holes in what we need to say in our document?
Frank: I suggested to a few people that we avoid the word "semantics" and use more specific terms where possible.
Mike: We need to acknowledge the role of semantics.
Katia: Maybe we could refer to relevant efforts.
Frank: We can use more specific concepts, which we do explain.
... In discovery, semantics is the information that is necessary to discover a service.
... If we say it that way, it reduces the fear of "semantics".
... In discovery, dbooth is already working on separating out and describing what's needed.
Katia: But the architecture doc is human readable.
Scribe: ACTION: dbooth to reference current semantics work (DAML-S, OWL-S) in discovery section
MikeC: Ok, we seem to be on the same wavelength here.
Hugo: This is work the WG did, and the idea is to make it available. The status section says that it does not represent WG consensus.
<mitrepauld> Does "service description" include "service lifecycle" state info (e.g., up/down)?
Zulah: Heather and I have an action item to work get all the Management stuff ready to publish, and we'll be talking next week about it.
Hugo: This is in publishable state. If zulah and heather are meeting next week, then they should make use of this.
Scribe: ACTION: Zulah and Heather to have proposed Management Note draft ready for WG to consider at next F2F
Frank: Reworked Management Model section.
... We painted a whole of management, and showed the relation to the rest of the architecture.
... It is focused on the idea that you may have a business relationship that allows a business partner to control some aspect of someone else's WS.
<mitrepauld> need to bring in the concept of delegation. An authority may delegate ...
Frank: The Management concepts and relationship is much cleaner and more obvious.
Zulah: It places management in context, but leaves details with WSDM TC.
Roger: It confuses me a little. I don't understand the emphasis on managing someone else's WS.
Frank: If you try to do it yourself, you have to manage the whole ball of wax. You start trying to manage a system.
... But from an interop perspective, where you can't look inside, you need to approach it differently.
Roger: THat makes sense. But it seems possible that others would have the same misapprehensive as me.
<mitrepauld> "back end agent" sounds a little too implementation-specific. why not "management agent"?
Frank: Also, dbooth and I discussed a small change to the def of "service".
... I also proposed a social model.
<mitrepauld> dbooth describing change to service def ...
<mchampion> dbooth: Noticed in doc that we sometimes define 'service' as a set of actions, and sometimes talk about
... service *performing* a set of actions.
<mitrepauld> set of actions vs thing that performs the actions ... logical inconsistency
<mchampion> This is a grammatical, possibly logical inconsistency.
<mitrepauld> small change to wording in concepts and relationships
... not def of "web service"
... only "service"
<mchampion> Proposed change to concepts/relationships definition of "service" [not re-opening the ur-Troutpond]
dbooth: Previous def said: "A service is a set of actions that form a coherent whole from the point of view of service providers and service requesters"
... Editorial change we made is "A service is a resource that is capable of performing a set of actions that forms a coherent whole ..."
<mitrepauld> katia - what about "agent"?
dbooth: No problem seen with "agent". The distinction betwen agent and service is described in the intro (sec 1.5.1), which will need to be changed slightly to reflect this change: http://dev.w3.org/cvsweb/%7Echeckout%7E/2002/ws/arch/wsa/wd-wsa-arch-review2.html#id61886
Scribe: (Various +1's to DBooth/Frank's new wording for the definition of "service")
<Daniel> D. Austin proposes that the social model is very helpful and should be part of the document
Katia: Not sure we should include Frank's social model
Roger: I like it
Zulah: Me too.
Roger: Maybe we should include it as a non-normative part.
Frank: I also added text on the 3 principles for Service Oriented Architecture.
<mitrepauld> Section 1.5 [Definition: A web service is ...] versus 188.8.131.52.2 "A service is a ...". IRC log above says "A Web service is a ..." - need to delete "web"
Eric: Some of the .PNG's are black.
<mitrepauld> IE "feature"
MikeC: THat's a known issue with IE.
dbooth: Hugo's makefile for the document will fix it.
MikeC: Paul Denning sent a list also
Roger: I can't address Paul's list because I dont' know what they all are. They're not annotated.
MikeC: what thoughts do people have?
Daniel: I'm trying to put it all into an ACCESS database, though i stripped out all the subjective info.
<mitrepauld> I had the list from an internal document bibliography, and saw that it included a few things that were not in Rogers list. Annotation would have taken longer, and I just wanted to get it out there.
Roger: That makes it useless.
<mitrepauld> Sorry for the heartburn
Paul: It would have taken me a long time to annotate it.
Daniel: I have 75 entries in the combined list.
<Daniel> paul: no heartburn, you did great work, it's jsut a lot of crazy data!
Roger: Paul separates items that I've lumped together.
MikeC: What should we do with this list? Do they align themselves with our taxonomy?
<mchampion> dbooth: what are we doing with this list? Mainly for Martin's action item to check how well these fit into our architecture
<mitrepauld> Mike mentioned "our taxonomy" ... what is this?
<mchampion> dbooth: Include the list as an appendix? That's an option
... katia: thought we agreed to do this
<mitrepauld> as opposed to the stack diagram?
... +1 to eric comment to map to the stack diagram
... rogers list at http://lists.w3.org/Archives/Public/www-ws-arch/2003Oct/0010.html
... Add categories under http://dmoz.org/Computers/Programming/Internet/Web_Services/ and add links to specs under pertinent topics
<mchampion> [various members expressed different opinions on whether to publish a list of everything, or apply selection criteria based on 'open standards' or whatever]
MikeC: If we use the list as a tool for our own purposes and avoid a lot of contentious debates. If we try to publish it as something from the WG, we'll need to agree on criteria for inclusion and selection and what to say about each.
Roger: It's important to list specs even if they aren't sanctioned by a standards body.
EricN: There's a huge difference between a proprietary spec and a spec produced by a standards body. It has a huge impact on a company's vulnerability to its use.
Scribe: (Discussion to continue on email)
... [Meeting adjourned]