See also: IRC log
<paulc> Regrets for today's meeting: Yakov S, Maryann (partial), Chris F (first hour), Frederick N, Toufic B, William H, Charlton B, Sergey B
<paulc> Regrets for Sergey is for only the second hour.
Paulc: several attendees have sent regrets. The agenda may need to be modified
scribe assignment for 16 May - Prasad
minutes approved
paulc: want members to register for F2F meeting
<Symon> aaaa is Symon Chang
abbie: meeting venue for F2F is still in flux, but will be in Ottawa, determined today
<scribe> ACTION: Abbie to update F2F info when available, including remote reachability status [recorded in http://www.w3.org/2007/05/02-ws-policy-minutes.html#action01]
<trackbot> Created ACTION-289 - Update F2F info when available, including remote reachability status [on Abbie Barbir - due 2007-05-09].
Abbie: describing plans for Internet access; static IP addresses still not resolved
Asir: updated scenarios sent out
... Primer document expected to be sent early next week.
topic 286: pending
288 - done
Asir: Update release 2 - minor changes are released
Paulc: dashboard summary out
Ashok: A namespace declaration for WS-A prefix was missing; we had to add that. The WS-Policy namespace declarations are slightly different - small changes
Paulc: are changes to interop scenarios needed for WS-A and WS-P changes?
<dorchard> For the Agenda Item 7, bug 4414: http://lists.w3.org/Archives/Public/public-ws-policy/2007May/0004.html
Paulc: Ashok, please open an issue on this.
Ashok: Second part does not need changes. It will look slightly different. I have sent out output results. I will reply to my my earlier message.
<scribe> ACTION: ashok to open an issue for the namespace problems in the WSDL 1.1 and potentially WSDL 2.0 external attachment tests [recorded in http://www.w3.org/2007/05/02-ws-policy-minutes.html#action02]
<trackbot> Created ACTION-290 - Open an issue for the namespace problems in the WSDL 2.0 external attachment tests [on Ashok Malhotra - due 2007-05-09].
<scribe> ACTION: asir to help ashok on AI 290 [recorded in http://www.w3.org/2007/05/02-ws-policy-minutes.html#action03]
<trackbot> Created ACTION-291 - Help ashok [on Asir Vedamuthu - due 2007-05-09].
<Zakim> asir, you wanted to ask what is the impact on WSDL 20 test files?
asir: two versions of external policy attachment, WSDL 1.1 and WSDL 2.0
Ashok: will address both.
<dorchard> http://lists.w3.org/Archives/Public/public-ws-policy/2007May/0004.html
Dorchard: feedback received from several sources
dorch: describes contribution
... should be able to be put in as is.
<monica> http://lists.w3.org/Archives/Public/public-ws-policy/2007May/0011.html
monica: This is moving in direction I suggested. The influence of mode, and understanding are not addressed in this contribution . We may need additional text.
Dorch: If there is missing information, we can add it later. We should try to get #44 closed, then add other.
Paulc: So Monica assents to the contribution?
Monica: yes.
Fabian: Mark it with option=true.
<monica> c/option/optional
Dave: I can do a table, or some other graphical way to describe it.
s/dave/dorch/
Monica: It may add to understandability.
Paulc: In summary, people support Dave's text, and he will add some additions.
<scribe> ACTION: Monica and dorchard to evaluate whether a new issue needs to be contributed for 4414 [recorded in http://www.w3.org/2007/05/02-ws-policy-minutes.html#action04]
<trackbot> Created ACTION-292 - And dorchard to evaluate whether a new issue needs to be contributed for 4414 [on Monica Martin - due 2007-05-09].
<monica> +1
<asir> Dave, you now have an editorial action to implement the resolution for issue 4414
<Fabian> what about the optional=true?
<Fabian> thanks
<paulc> Proposal to resolve 4414:
<paulc> a) text for message April 0052
<paulc> b) text from message May 004
<paulc> c) add a table to show different cases covered and results
<paulc> d) implement Fabian's suggestion to replace "one alternative with and one alternative without" with text using wsp:Optional=true
Paulc: This resolves 4414.
<fsasaki> RESOLUTION: above list a-d resolves 4414
<asir> Asir's amendment is http://lists.w3.org/Archives/Public/public-ws-policy/2007May/0004.html
<paulc> a) Issue 4479: Editorial, note is obscure or unclear
<asir> Bob's +1 is at http://lists.w3.org/Archives/Public/public-ws-policy/2007May/0013.html
Asir: I folded in the example. that's the only change.
bob: It is more understandable, not conveying new or unusual information.
resolution: contribution resolves 4479
<dorchard> I have to drop off for about 20 minutes, I'll try to be back for the "oh my" discussion.
<asir> proposal is at http://lists.w3.org/Archives/Public/public-ws-policy/2007May/0005.html
<paulc> Correct link and solution is : http://lists.w3.org/Archives/Public/public-ws-policy/2007May/0005.html
contribution is actually may/0005.
<asir> ACTION: Asir to close out issue 4479 [recorded in http://www.w3.org/2007/05/02-ws-policy-minutes.html#action05]
<trackbot> Created ACTION-293 - Close out issue 4479 [on Asir Vedamuthu - due 2007-05-09].
<paulc> rssagent, where am I?
Paulc: skip item b
<paulc> Sets and bags: http://lists.w3.org/Archives/Public/public-ws-policy/2007May/0001.html
Asir: I responded to this yesterday.
... I answered the two questions.
<paulc> David's reply to Asir's reply: http://lists.w3.org/Archives/Public/public-ws-policy/2007May/0020.html
Paulc: Look at message 20.
Asir: First question: the answer is yes.
... Second question: special meaning for duplicates? Answer is no.
... Duplicates need to be taken out? Difficult on developers.
<prasad> I relate to collection better than a bag
Ashok: "Bag" should be used here.
Asir: Not support use of bag.
Paulc: The term collection appears several times.
<paulc> A policy is a collection of policy alternatives. A policy alternative is a collection of policy assertions
Paulc: Ashok - do you want to use the term "bag" to convey the meaning of unordered group?
Ashok: yes
<TRutt__> Issue 4479: Editorial, note is obscure or unclear
<TRutt__> http://lists.w3.org/Archives/Public/public-ws-policy/2007Apr/0081.html
<TRutt__> Issue 4479: Editorial, note is obscure or unclear
<TRutt__> http://lists.w3.org/Archives/Public/public-ws-policy/2007Apr/0081.html
MOnica: maybe " a collection in the context of this document is:"
Ashok: what is definition?
Paulc: "the items in a collection in this specification are unordered and may contain duplicates"
Ashok: "may contain duplicates"
<scribe> ACTION: ashok to open a CR issue with text to define "collection" [recorded in http://www.w3.org/2007/05/02-ws-policy-minutes.html#action06]
<trackbot> Created ACTION-294 - Open a CR issue with text to define \"collection\" [on Ashok Malhotra - due 2007-05-09].
Paulc: Ashok, point to David Hull's message and identify the usage of "collection" for use as a hyperlink
<paulc> http://lists.w3.org/Archives/Public/public-ws-policy/2007May/0002.html
<paulc> Asir: http://lists.w3.org/Archives/Public/public-ws-policy/2007May/0006.html
Asir: Message 0006 is my response
... I suggested David Hull open a CR issue with this point.
<paulc> David's reply: http://lists.w3.org/Archives/Public/public-ws-policy/2007May/0019.html
Asir: For the second point, I walked through
the handling of the empty policy alternatives. His response is in message
0019.
... describes the message exchange
<scribe> ACTION: Asir to address the first item in message 2007May/0002. [recorded in http://www.w3.org/2007/05/02-ws-policy-minutes.html#action07]
<trackbot> Created ACTION-295 - Address the first item in message 2007May/0002. [on Asir Vedamuthu - due 2007-05-09].
Asir: I will open a CR issue.
<asir> asir to open a CR issue based on item 1 in David Hull's e-mail titled 'Editorial Issues'
<paulc> "Normal form" http://lists.w3.org/Archives/Public/public-ws-policy/2007May/0000.html
Paulc: We need someone to respond to the email
thread.
... Issue 3621 was this topic.
Asir: He was asking fro formal semantics, using his own notation, which appeared to produce the same results as the spec. That was his conclusion.
Paulc: It is just a comment, not a change request.
<asir> Ashok didn't say what is unclear
Paulc: I will try to respond to David, pointing him to 3621
<asir> David Hull's statement is - "I may well have missed some subtleties, but this approach appears to produce the same results as the spec."
Fsasaki: Don't see a need to do this in formal semantics, since nothing is broken.
<monica> new from Hull: http://lists.w3.org/Archives/Public/public-ws-policy/2007May/0024.html
Chris Ferris: Describing message 0003.
<paulc> policy vocabulary, will not be applied, oh my!
<paulc> http://lists.w3.org/Archives/Public/public-ws-policy/2007May/0003.html
Chris Ferris: The proposal includes a change-markup version. It describes handling the absence of an assertion, among other clarifications.
Monica: How does it clear things up? we need to
be explicit.
... Not clear if absence means negation, based on the text.
Chris: We want to be clear that absence is negation.
Monica: Is that the consensus of the WG?
Paulc: Monica, do you have text?
Monica: not yet.
Dorchard: We should be either fish or fowl,
either fully in or fully out.
... The clarification of absence as negation is still unclear. We need a
clear use case and I have not seen it yet. It's not clear if it is in
scope.
... Is Chris saying RM needs "absence as negation"?
Chris: FOr effective interchange, it is needed.
Ashok: We need to look at it as two quite different questions: 1 - is vocabulary-based negation to be removed? I think it should be removed.
<Fabian> sorry, testing, my network connection is close to unusable right now
Ashok: 2- do we require negation? What form shold it take? We have not spent enough time on this. We need use cases.
<danroth> An alternative with zero assertions indicates no behaviors. An alternative with one or more assertions indicates behaviors implied by those, and only those assertions.
Danroth: Agree with Ashok to remove negation. I
will paste in IRC for second point.
... Need to know all behaviors relative to interop. No other behavior should
be implied.
... I do what I say, and nothing else.
Chris: I don't understand the limits of "nothing else." It's close to absence as negation.
<cferris> yes
<monica> http://www.oasis-open.org/apps/org/workgroup/ws-rx/download.php/23373/wsrmp-1.1-spec-cd-09.pdf
<cferris> I am here
Monica: Look at the links I am posting.
<monica> Any messages, regardless of whether they have an attached Message Policy Subject RM policy assertion, MAY be sent to that endpoint using WS-RM. Additionally, the receiving endpoint MUST NOT reject any message belonging to a Sequence, simply because there was no Message Policy Subject RM policy assertion attached to that message. There might be certain RM implementations that are incapable of applying RM Quality of Service (QoS) semantics on a per-message b
<asir> in other words, what Dan mentioned is ... no other behaviors are implied
Monica: This doesn't seem to imply absence is
negation.
... text is from 335 in RMP spec.
Chris: This refers to attaching to an endpoint, not at the message level.
Monica: Describing case.
Chris: Two different levels of policy attachment.
Paulc: a different level in the WSDL.
... This may not be the best example to discuss the issue.
<danroth> The RM text appears to be consistent to with the framework text that no other beahviors are implied
Prasad: General support Chris' proposal. Is it too complex, requiring a policy to define all behavior? It is getting like WSDL now.
Chris: To make it work, you have to assume the published policy is complete.
<danroth> +1 to Chris. You have to completely and accurately author your policy
Chris: Either it is complete or it is a complete waste of time.
<Zakim> dorchard, you wanted to respond about RM, AIN and negation
Dorchard: Trying to understand it. If we require absence as negation, it is confusing. Example using RM.
<dorchard> My example is: Service has an endpoint that supports RM and another that doesn't.
<monica> +1 to prasad, ashok, and to dave - no position on negation or not - but the all knowing aspect.
<dorchard> Client supports RM and not RM. If it wants to use RM, it will intersect with the endpoint that says RM.
<dorchard> If it doesn't want to use RM, it will intersect with the endpoint that doesn't use RM.
<dorchard> If the client decides to use RM with the endpoint that doesn't support RM, it's ignoring the intersection result it got!
Ashok: There are situations where you don't want all assertions. It is difficult.
<monica> +1 to bob
<cferris> in response to Ashok's and Prasad's point... I am not saying that you have to know about everything... I am rather saying that when a policy is published, that it is a full accounting of that which it knows. Hence, if an assertion is not present, one can infer that it is either not "supported" or not wanted
Bob: The absence of an assertion is wholely domain-dependent . Absence-as-negation needs to be under the control of the domain.
<asir> There are three items here ... a) drop the "will not be applied" statement, b) drop related vocabulary definitions and c) Chris' request to add something .. there is consensus on items a) and b) and discussion on c) .. A way to move forward is to process a) and b) and continue discussion on c)
Asir: The discussion has reached consensus on two items, and we should concentrate on the absence-as-negation issue.
Bob: Domain-specific definition of absence-as-negation is needed.
<danroth> If the semantics of an absent assertion can be a defined by the assertion, then it will be impossible to author policies. You will need to understand what the absence of every assertion in the world means in order to understand your policy.
Trutt: Absence is not always defined in every domain. Negation is not always obvious in meaning.
Chris: Support Asir's proposal. I don't propose
to remove a sentence, but to replace it.
... Don't support that part.
<TRutt__> For example wsrmp:InOrder means that the receiver will process to ensure ordered delivery. DOes its negation mean that the messages must not be delivered in order?
<Zakim> dorchard, you wanted to respond to Asir's suggestion. I disagree on a) and b), sorry.
Dorchard: No consensus on Asir's proposal of dropping vocabulary definitions. They are needed along with absence-as-negation.
<monica> where is the current use case for absence=negation? Articulate the use case.
<TRutt__> Also nested assertions have a context (parent assertion type) to define what missing means, however top level assertionas have glogal context only
<danroth> NOBI = No Other Behavior is Implied
Dorchard: would like to post response to mail list and try to nail the issue.
<asir> NOBI
<paulc> NOBI
<asir> NOBI is different from AIN
<cferris> regarding daveo's point about the RM use case I gave, I am not implying that RM ***needs* AIN, I am stating that if a policy does not say something, that you need a framework within which to understand what that means
Danroth: Policy should assert behaviors, not
imply behaviors by absence. NOBI is an acronym.
... No Other Behavior is Implied.
<monica> Can Daniel put on email what he means by NOBI and differentiate from AIN?
<monica> so we understand?
<danroth> yeah, I'll send out a mail
<monica> thanks
Adjourned 2:01