See also: IRC log IRC log
Present: Mike, Roger, Zulah, Frank, David, Hugo, Jeff, Bijan_(observing), Philippe_Le_Hégaret_(observing)
Regrets:
Chair: Mike
Scribe: Hugo
Scribe: Agenda: http://lists.w3.org/Archives/Member/w3c-ws-arch/2003Oct/0095.html
Mike: Tomorrow, we'll be discussing the OWL formalization first
thing
... Katia should be calling in
<mitrepauld> http://lists.w3.org/Archives/Member/w3c-ws-arch/2003Oct/0095.html
... Some of 0900-1100 time slot Thu to cover Discovery.
<bijan> MIND lab's message reviewing the OWL formalization: http://lists.w3.org/Archives/Public/www-ws-arch/2003Nov/0002.html
<mitrepauld> meta agenda: go thru whole doc. Rough sense of what will be there. Contentious items need decision.
Mike: we really want to leave this meeting with a sense of what is
going to be in this document at the end of January
... we are going to have to deal with potentially contentious matters one way
or another: vote on it, remove it, reword it in a non-contentious way, ...
Roger: what about reliability?
Mike: we'll do it Thursday afternoon
... we cannot go through the spec paragraph by paragraph
... issues that haven't come up before Friday noon will be considered
somewhat out of scope
<mitrepauld> baseline for review is Editors' copy $Date:
2003/11/05 23:52:15 $ @@ @@@ 2003 at http://dev.w3.org/cvsweb/~checkout~/2002/ws/arch/wsa/wd-wsa-arch-review2.html
... http://dev.w3.org/cvsweb/~checkout~/2002/ws/arch/wsa/wd-wsa-arch-review2.xml
has date 2003/11/05 16:59:14
Scribe: We are discussing: http://dev.w3.org/cvsweb/~checkout~/2002/ws/arch/wsa/wd-wsa-arch-review2.html?rev=1.75&content-type=text/html;%20charset=iso-8859-1
Scribe: We have removed the editorial note which isn't necessary anymore
Zulah: what is "frameworks"?
Scribe: The group is unsure
<mitrepauld> I vote leave it.
... Not too strong on this either.
... cut it.
... The WSA promotes interoperability through the identification of
compatible protocols.
Mike: we should remove the above sentence
Scribe: Decision is to remove it
Roger: I think that the audience of this document is more
restricted than what we are advertizing
... I think that people interested in this are people who are making
decisions in putting resources into a Web services project
<mitrepauld> for people making strategic decisions regarding web services and related technologies (OWL/Semantic Web, REST, SOA)
Frank: a role for the document is to predict what you would expect to see in this domain
Zulah: what we have here is a conceptual architecture; I think that it is benefitial to spec writers
<mitrepauld> I agree with zulah.
Zulah: but this is not restricted to those people
... listing the set of people who would be interested in this is IMO
difficult and useless
<mitrepauld> Don't want authors of future WS specs to deviate
from WSA, or at least (if they deviate) to be able to show how it relates to
WSA.
... Provides a basis for diverse organizations to work together.
Scribe: David puts a sentence reflecting Zulah's point in the document
<mitrepauld> common conceptual model, which provides ...
Roger: I think that the concept / relationship difference is blurred here
Decision: remove everything but the first sentence from the first paragraph
<mitrepauld> In a telecon we said something about a lot of people caring about the Stakeholder, and techies looking at the concepts. So should Stakeholders be first so the majority of readers do not have to skip over Concepts section?
Scribe: More changes have been mode by David to the first paragraph
<mitrepauld> Do we need to align structure/outline of WSA with TAG Webarch?
Scribe: Wordsmithing a statement about conformance
Jeff: I don't think that it's useful to talk about conformance here
<mitrepauld> I have to leave for a few minutes; I'll dial in again when I come back.
Scribe: Removed the reference to list of technologies
Scribe: We may remove the section if we don't use them
... An editorial note has been added
Roger: I would like to replace the word "concrete entity" by "program"
David: this may be ambiguous
Mike: is 1.5.3 consistent with what the WSDWG thinks?
David: yes, it is
Roger: the view of semantics in 1.5.4 is a little weird
... semantics isn't behavior
Zulah: it is common in computer science litterature
... but I think that it is too narrow here
Roger: it's not the behavior, it's the expectation
Decision: first sentence of 1.5.4 changed into "The semantics of a Web service is the shared expectation about the behavior of the service [..]"
Zulah: we should strike the second paragraph
Scribe: Done
Roger: I would like to add a step 3 to 1.5.5: humans use the service
Mike: not necessarily
Roger: somebody needs to run the service
Zulah: what's the purpose of this section?
Mike: the goal is to remind people that humans are still in the picture
David: I think that it's ueful to show where the work has been shifted from the human to the machine
Frank: there are scenarios where automated discovery is needed
Mike: we need something saying that today the semantics are basically done by humans, and that in the future we expect more and more automation
Zulah: I think that this paragraph gives a wrong idea about the architecture
David: I don't understand how this could be misunderstood
Zulah: because it's right next to semantics
Hugo: how about moving the paragraph up in section 1.5?
Zulah: I wouldn't mind having a section above talking about the
role of humans in integration
... one of the roles is to remove the role of humans in the picture
Roger: I disagree; David's view is that it is never removed, but moved
Zulah: the purpose is to integrate
... it would be nice if the integration was less brutal
Scribe: ACTION: Mike to initiate a discussion about section 1.5.5 The Role of Humans to resolve the difference of opinions
David: the purpose of the section is to illustrate the context in which Web services are
Scribe: -- Break --
... ACTION- 1
... an adequate wording has been agreed upon
Mike: section 1.5.6 is too long
David: I agree
Mike: it should probably be moved to the stakeholders section
Scribe: ACTION: David to refactor "1.5.6 The Process of Engaging a
Web Service"
... the paragraph under figure one should be kept, and the rest should be
moved
Mike: some text about a part being automated and another being manual should be added there too
Scribe: ACTION: David and Roger to clarify step 3 of section "1.5.6 The Process of Engaging a Web Service"
Mike: I don't think that the goals listed in "1.6.1
Interoperability Architecture" have been met
... I think that we have learnt things, but I don't think that we have
improved much Web services interoperability
Frank: SOA is an interaction architecture
Roger: I don't like listing the non-goals: I think that it's
gratuitous and it's a duplication
... since it is an introduction to the rest, removing the duplication from
the requirements document may do the trick
Hugo: then we should remove the title
Decision:
Scribe: - first paragraph moved to "purpose" section
... - "architectural style" -> "SOA"
... - "interoperability architecture" section nuked
Roger: what isn't an SOA?
David: parallel computing
... but does it matter?
... is discovery a part of SOA?
Zulah: we say that our architecture is an SOA
... and ours allows us to do automatic discovery
Scribe: David adds an item about discovery
Roger: I want to add that a Web service doesn't necessarily use automatic discovery
Frank: the CORBA paragraph aims at saying that CORBA is much more integrated
Mike: we need some text about the difference between SOA and distributed systems
<dbooth> ACTION: MikeC to propose text to replace these next three paragraph, to explain the difference between an SOA and a distributed system.
Scribe: ACTION 4 = MikeC to propose text to replace these next
three paragraphs (and the diagram), to explain the difference between an SOA
and a distributed system.
... ACTION: Hugo to clean up "1.6.3 SOA and REST architectures"
Zulah: this section is difficult to read
Scribe: http://www.w3.org/TR/webarch/
... http://internet.conveyor.com/RESTwiki/moin.cgi/WhereRestStopsAndHttpStarts
... Discussion about how to characterize REST and HTTP
David: do we need to say anything about that?
Scribe: Discussions around visibility
... The group feels that capturing a definition of visibility in the
document
<dbooth> ACTION: Frank to move these next two paragraphs to the Semantics stakeholder section
Scribe: ACTION 6 = Frank to move the last two paragraphs of "1.6.3
SOA and REST architectures" to the Semantics stakeholder section
... ACTION- Hugo
... ACTION- 5
... ADJOURNED