See also: IRC log
previous minutes at http://www.w3.org/2007/01/23-ws-cg-minutes.html approved
Philippe: I moved status report item after xml protocol charter review, since Paul can only stay for 1/2 hour
ACTION: plh to research IPR implications of a Core group (DONE) [recorded in http://www.w3.org/2007/02/06-ws-cg-minutes.html#action01]
Philippe: we should put in the charter that the
group is only allowed to do a Proposed rec without new features
... if we do a new feature, we'd need a new WG. Paulc. are you ok with
that?
Paul: Good enough, you did your AI. I'll check with laywers
Philippe: Didn't ask Rigo about it, only talked to Ian Jacobs. We can come back to this question once we have a more concrete proposal on the table.
Philippe: question is: what is the best NS you
can get today?
... policy has decided to use new /ns convention without the year
... the drawback is: if there is a substantial change during CR, people will
be unhappy if you change the semantics of the namespace
<pauld> likes the year
Paul: the NS for is a little bit easier to remember, but you might end up using "ns policy 2nd edition" NS
Chris: so we have two choices: stick with /ns
and might be forced to change it during CR
... (if we go to a new version: what do you do? only forward / backward
compatible changes/?)
... I don't think there is a right answer
Philippe: agreed there is no right answer
... the problem is the same as with the dated NS
... To my best knowledge, for V.Next, WGs are just adding new stuff to their
namespace and try not to break it
Paul: so policy could have choosen /2007/03 ,
or /ns form
... but in any case the best policy is not to change it during CR
Philippe: unless you do substantial changes you don't change, right?
Paul: want to distinguish between going back
and forward
... so e.g. if you make a change to a feature and have to go back to WD, you
can't use the namespace anymore
Jacek: are you objecting against the /ns namespace policy?
Paul: no, my point is: if you go backwards during CR you have "burned" the /ns namespace
Jonathan: it occurs like wsdl would have a /ns/wsdl20
Philippe: it is up to the WG to decide what
they want
... another way to avoid a different version of the /ns is to say: we would
keep the same namespace in any case
... but some WG participants might disagree with such approach if they start
implementing the specification
Jonathan: for wsdl, you could do /ns/wsdl2
... but I would challenge the idea that it is bad to change the NS between CR
> PR
... it depends very much on the implementers community
... I thought about this recently, and maybe WSDL could make the change
... we did a lot of changes
Philippe: the changes were marked at risk btw
Jonathan: still there was a lot of changes
Jacek: agree with most of what jonathan said
... I'm against /ns/wsdl2
... difference between wsdl2.1 or wsdl2.2 is not good, I'd like to have
/ns/wsdl
... remove of MEPs would not be a reason to do the change of the NS
Jonathan: there was a lot of changes
Jacek: do we need to go back to LC?
Philippe: don't think so, there was a lot of
clarifications
... but let's not discuss this now
Jonathan: I'm in favor of using the /ns version for wsdl and want to bring that to the WG
Paul: I concur with Philippe: It depends on the
WG and what they want to achieve
... agree with jonathan to bring this to his WG
... jonathan is looking back from "after CR" so that's an easier position to
be in.
Jonathan: at entering CR I thought there might
be changes which are close to significant
... maybe different to policy
Paul: in xquery there was a change of data types without a new NS, and the WG agreed
Philippe: In any case, it is important to describe what you want to do in the NS document when you move to CR. people are looking at that document for implementation expectation and might call you on it if you don't say anything.
<cferris> +1 to plh that you document the namespace policy in the namespace document. ws-policy did this
Jonathan: Jacek said it would be nice if namespaces would be "together" like various WSDL versions
<cferris> http://www.w3.org/2006/07/ws-policy/
Jonathan: we can decide independently but it would be good to know what WGs do
<cferris> It is the intent of the W3C
Web Services Policy Working Group that the Web Services Policy 1.5 -
Framework and Web Services Policy 1.5 - Attachment XML namespace URI will not
change arbitrarily with each subsequent revision of the corresponding XML
Schema documents but rather change only when a subsequent revision, published
as a WD, CR or PR draft results in non-backwardly compatible changes from a
previously published WD, CR or PR draft of the specification.
Jacek: couldn't WGs say that they expect to change the /ns during CR?
Philippe: needs to be discussed by each WG. W3C
introduced /ns because of feedback of some WGs.
... the dated version is still available and W3C doesn't give any preference
to one or the other.
Jonathan: what will your WG do, Jacek?
Jacek: prefer to do what WSDL does
Philippe: did not talk to policy folks before
sending message on interop event
... idea is to have an interop event where everybody can participate
Paul: we are not in CR yet, will decide that
tomorrow
... I want to get out of CR via the March / May f2f (without consulting
Chris)
... we could use the second f2f for interop
... the May meeting is scheduled to be in Ottawa
Philippe: in WSDL and Addressing, the interop events were before / after the actual f2f, not during the actual f2f.
Paul: would other chairs find this useful?
Bob: does May avoid XTech and WWW 2007 / AC meeting?
Paul: we looked at WWW 2007 / AC, we did not
look at XTech
... we planned for 23,24, 25 May
... to make clear: we ask "you come to the interop event and bring the
implementation of your policy assertion"?
Bob: I guess that is the expectation
Chris: as XML Protocol chair we have a draft
that is 4 days old
... I don't think we would be ready. The policy WG would need to develop
tests for a May interop event
Philippe: for the f2f in May, are you including the interop idea?
Paul: we asked the hosts of both f2fs if there
are rooms for interop tests
... beyond that there are no detailed plans
... the WG expects that there will be interop tests during the f2f
Bob: I'll check what the developer overlap is between implementations WS-Policy and Addressing policy assertions
Philippe: For Policy, you did not discuss in the WG how you want to do the testing, correct?
Paul: true. However, IBM and MS submitted interop scenarios
Philippe: might be difficult for people to integrate their own tests into that
Paul: we don't need to provide that to leave CR, right?
Philippe: correct. On the interop stuff: is it "just" test cases?
Chris: yes
Philippe: test cases is easy, the framework to
check results is the hard part
Chris: not sure if there will be a framework for testing
Paul: last time we had an interop event without
a framework, and it worked
... having a framework or not is quite subjective
ACTION: bob to ask addressing for test development for policy assertions [recorded in http://www.w3.org/2007/02/06-ws-cg-minutes.html#action02]
ACTION: cferris to ask xml protocol for test development for policy assertions [recorded in http://www.w3.org/2007/02/06-ws-cg-minutes.html#action03]
ACTION: paulc and cferris to ask policy WG about CR exit criteria [recorded in http://www.w3.org/2007/02/06-ws-cg-minutes.html#action04]
[Paul Cotton leaves the call]
Chris: I asked everyone to look at Jonathan's note to have a discussion
Philippe: does the WG just want to take the member submission and put it to REC, or do more functionality?
Chris: we will take the submission as a
starting point and will do issue resolution against it
... there are some things the WG may want to address, but I did not hear a
desire to extend the functionality
... there is not much to extend
Jonathan: the granularity could be extended
... Canon is interested in a high level of granularity
Philippe: XML Protocol has to decide whether
they want to extend the granularity
... the WG should discuss it and put it into the charter
Jonathan: the charter is not clear if the document will be feature or schedule driven
Chris: the WG does not want to make that their lifes project
Jonathan: the WSDL WG had different answers to
that question over time
... would have been better to have it explicit in the charter from the
start
... would recommend to put a stake in the ground from the beginning
Philippe: IMO, it is important for this WG to be schedule driven and the charter should emphasize this. We don't need an MTOM policy assertion in two years.
[Paul Downey leaves the call]
Philippe: Anything to report to the CG?
Chris: we intend to have a poll tomorrow about transition to candidate rec
Jacek: semantic annotation might be able to become a REC together with WSDL. What is the timeline for WSDL?
Jonathan: we have about 10 issues open, interop testing is ongoing. So in best case, we could move in 2-3 weeks.
Jacek: that is maybe outside our reach
Jonathan: you never know, we could be stuck on an issue.
Bob: addressing metadata is in LC, we are
looking for feedback
... we hope to go to CR immediately
Philippe: policy, description, ws-rx ought to give feedback on that document
Chris: about XML Protocol: got FPWD of one-way
MEP
... asked addressing and WSDL to look at that
Jonathan: WSDL WG should provide comments by the end of the week
Chris: after that we will target last call, around March
Philippe: feb 20, any regrets?
[none heard]
Philippe: Paul Downey said this time is difficult. Would people be willing to change?
no objection (in general)
Philippe: I'll put it on the agenda for the Feb
20 call
... If at all possible, I'm guessing the only option is to move it to a
different day at the same time
Jacek: might be difficult for me. This call is easy since it's just after the SAWSDL one.