WS Policy WG
6 Jun 2007

Administrivia

Secratary for June 13th is Toufic, June 20th is William Henry

<charlton> To mute yourself, press 61#, unmuting is done py pressing 60#, or use IRC

<maryann> yes

Schedule Review

this is a reminder of what's coming next (based on Ottawa F2F discussion

[Chris walks through the schedule]

1) May 30th [DONE][1] - Editors to deliver the updated Framework and Attachment drafts

2) June 5th - Refresh the Framework & Attachment CR drafts and Primer WD

3) June 20th - Editors to deliver the updated Framework and Attachment drafts

[WSO2/Jonathan is looking into when they'd report their last two unit test results]

4) June 27th - WG decision to request the director to advance the Framework and Attachment drafts to PR

5) July 2nd - Co-chairs to create the request to advance the Framework and Attachment drafts to PR

[Director's call is scheduled for July 2nd]

6) July first week - Co-chairs and staff contact to meet with the director

7) July - Advance to Proposed Recommendation

8) July Dublin F2F - Process Primer and Guidelines doc issues, V.Next issues

9) August - Advance to Recommendation

10) August, last two weeks - Break

[We may take additional weeks of depending on where the primer and guidelines doc are]

11) September-October - Post Recommendation work, Process Primer and Guidelines doc issues

12) November 9-10 Cambridge F2F - Post Recommendation work, Finalize Primer and Guidelines documents, V.Next and Next Step Considerations

Review and approval of WG minutes

RESOLUTION: adopted minutes from the Ottawa F2F

http://www.w3.org/2007/05/23-ws-policy-minutes.html

http://www.w3.org/2007/05/24-ws-policy-minutes.html

Future WG meetings

Chris - can you update the schedule on the admin page

Felix: Done!

Chris: if you haven't registered for the Dublin, please register ASAP

Editors' Report

Asir: delivered Framework and Attachment drafts, Guidelines and Scenarios are a work in-progress

<maryann> ongoing

Review action items

Action 279 is ongoing

Action 286 is ongoing, ETA - next week

Action 300 is pending

Action 301 is done

Action 302 is done http://lists.w3.org/Archives/Public/public-ws-policy/2007May/0276.html

<maryann> just a sec

Action 303 is pending, ETA - next week

Action 304 is pending

<maryann> yes

Action 305 is pending, ETA - next week

<maryann> zakim please unmute me

Action 306 is pending

WS-Policy Call for Implementations

Interop results are updated - http://dev.w3.org/cvsweb/~checkout~/2006/ws/policy/interop/results/dashboard-summary.html

We are waiting for one feature to be tested by another impl - this is WSO2 - they are actively engaged in working on this feature

<cferris> http://www.w3.org/Bugs/Public/show_bug.cgi?id=4582

ISSUE 4582 test cases need coverage of nested policy for lax interse...

<fsasaki> asir: looked at the test case, found some minor issues

<fsasaki> .. expected outcome is in partial normal form, not full normal form

<fsasaki> .. some naming issues for files

<fsasaki> .. at f2f we said these test cases would be optional and marked as round 5

<fsasaki> .. not sure if we can run them, want to see what other participants will do

<fsasaki> chris: other means IBM?

<fsasaki> asir: still looking into it

<fsasaki> asir: half of the output is in normal form, the other half in compact form, I'll send a mail about that

<cferris> ACTION: Asir to send updates to proposed new test cases (round 5) regarding normal form of output [recorded in http://www.w3.org/2007/06/06-ws-policy-minutes.html#action01]

<trackbot> Created ACTION-307 - Send updates to proposed new test cases (round 5) regarding normal form of output [on Asir Vedamuthu - due 2007-06-13].

<ArnaudM> Thank you ;-)

Chris: thanks everyone for participating at the Ottawa interop event and showcasing all those greens

Liaison items

Liaison is ongoing with WS-A .. will be until WS-A exits CR

Chris: b) should have been deleted

[Bug 4561] clarification of domain-specific processing

http://lists.w3.org/Archives/Public/public-ws-policy/2007May/0253.html

<cferris> http://lists.w3.org/Archives/Public/public-ws-policy/2007May/0253.html

<cferris> http://lists.w3.org/Archives/Public/public-ws-policy/2007Jun/0017.html

<fsasaki> asir: mail about continuation from f2f

<fsasaki> .. at f2f we discussed about domain and nested policies

<fsasaki> .. if two alternatives are compatible, there is no issue

<fsasaki> .. if two alternatives are not compatible (qnames different), there is no way to override that

<fsasaki> .. this approach returns never false negatives

<fsasaki> .. if there would be false negatives a requester could not override the result of intersection

<fsasaki> .. I made some clarification proposals in the mail

<fsasaki> (asir reading the proposals)

<fsasaki> chris: what is the point of emphasizing false negatives?

<fsasaki> asir: if policy intersection would produce false negatives you could not override that

<fsasaki> chris: when did false negatives up?

<fsasaki> asir: I think at the f2f

<fsasaki> chris: so scenario is: domain independend intersection first, other later?

<fsasaki> .. that would be bad? Either a domain has it's own thing or not?

<fsasaki> asir: qname won't tell you if domain specific intersection is required or not

<fsasaki> chris: is domain intersection completely distinct from domain intersection?

<fsasaki> asir: intersection algorithm gives parameters as a means to override the generic intersection

<fsasaki> chris: what is other part of domain-specific processing besides parameters?

<fsasaki> .. is domain-processing limited to assertion parameters?

<fsasaki> asir: policy intersection has extensibility points, here it delegates to the domain

<fsasaki> chris: is domain-specific processing for compatibility testing limited to parameters?

<fsasaki> asir: yes

<fsasaki> chris: the text proposal does not say that

<whenry> Can we have a practical example of when false negitives might occur?

<fsasaki> .. "part of" assumes that there is another part

<fsasaki> ashok: asir, you wrote first you do domain-independent intersection

<fsasaki> .. depending on the results you might apply domain-specific rules

<fsasaki> .. I don't disagree, but we should make that clearer: first you do this, later that

<fsasaki> asir: that is not a requirement

<fsasaki> .. it would work either way

<fsasaki> ashok: explain why

<fsasaki> .. for domain-specific processing, what are you trying to catch?

<cferris> Mechanisms for determining assertion parameter compatibility are not part of the domain-independent policy intersection. Determining whether two policy assertions of the same type are compatible MAY involve domain-specific processing for purposes of determining assertion parameter compatibility. If a domain-specific intersection processing algorithm is required this will be known from the QName of the specific assertion type involved in the policy alternat

<cferris> As a first approximation, the following algorithm is defined that approximates compatibility in a domain-independent manner:

<fsasaki> asir: crawler is unaware of domain specific rules

<fsasaki> .. a different policy processor might use domain-specific rules

<fsasaki> ashok: how do you do domain-specific intersection without seeing what is compatible?

<fsasaki> asir: the rules in sec. 4.5 make that clear

<fsasaki> .. that is spelled out

<maryann> q

<maryann> q

<maryann> =

<fsasaki> .. between 2 and 3, you see that it is delegated to the domain

<fsasaki> maryann: asir, if you use the language about "extensibility points" you mentioned, that might help with ashoks question

<fsasaki> chris: agree

<fsasaki> .. asir describes an extensibility point an implementation may choose

<fsasaki> .. that need not to be part of the domain-independent framework

<fsasaki> asir: chris, could you type the other clarification?

<fsasaki> chris: let's take it to mail

<cferris> so, domain-independent intersection MAY be extended to include domain-specific processing. However, regardless of whether an assertion's QName indicates domain-specific processing, an implementation of the domain-independent intersection need not apply the domain-specific processing

<fsasaki> asir: good statement

<fsasaki> chris: I'll take that to mail and add it to the paragraph above

<cferris> ACTION: Chris to follow-up Asir's email http://lists.w3.org/Archives/Public/public-ws-policy/2007Jun/0017.html with proposal based on this discussion [recorded in http://www.w3.org/2007/06/06-ws-policy-minutes.html#action02]

<trackbot> Created ACTION-308 - Follow-up Asir\'s email http://lists.w3.org/Archives/Public/public-ws-policy/2007Jun/0017.html with proposal based on this discussion [on Christopher Ferris - due 2007-06-13].

<fsasaki> asir: what is pending now is wordsmithing for sec 4.5?

<fsasaki> chris: yes

<monica> chris i was on the queue

<monica> will go to email

[Bug 4583] Duplicate assertions in an alternative

<fsasaki> ashok summarizes the discussion

<fsasaki> see thread at http://lists.w3.org/Archives/Public/public-ws-policy/2007Jun/0002.html

<fsasaki> asir: ashok suggests to take Dale's sentence and my amendment and put to sec. 4.5?

<fsasaki> ashok: I want more than that, your wording modified Dale's wording

<fsasaki> .. you speak about assertions that don't have parameters

<fsasaki> .. for removing dublicates from intersection results, you don't need to do that

<fsasaki> asir: did not see proposals to remove anything

<fsasaki> .. Dale said "if there are multiple assertions, it is the same as a single one, looking at the assertion type"

<fsasaki> .. how would you handle parameters in nested assertions?

<fsasaki> ashok: we are not talking about parameters anywhere

<fsasaki> .. I can write a proposal

<cferris> ACTION: Ashok to provide a counter proposal that adds to the sentence proposed by Dale [recorded in http://www.w3.org/2007/06/06-ws-policy-minutes.html#action03]

<trackbot> Created ACTION-309 - Provide a counter proposal that adds to the sentence proposed by Dale [on Ashok Malhotra - due 2007-06-13].

<fsasaki> asir: so you are ok with Dale's sentence, plus the amendent clarification? You want two sentences added, and you are working on the second sentence?

<fsasaki> ashok: yes

[Bug 4592] incorrect reference content

Chris walksthrough the minor editorial issue

Resolution: adopts the proposal in bug 4592

<cferris> RESOLUTION: http://www.w3.org/Bugs/Public/show_bug.cgi?id=4592 resolved as proposed

[Chris is closing the bug]

Comments on "normal form" of policy expressions

Chris: do we have an issue?

<fsasaki> asir: this thread is related to 4598

<cferris> http://www.w3.org/Bugs/Public/show_bug.cgi?id=4598

<fsasaki> .. has been opened recently

<fsasaki> .. if we close 4598, we can close the thread

<fsasaki> .. david hull had two questions which we explained, I think we are done

<fsasaki> ashok: not quite right, he wants some better wording

<fsasaki> asir: the last sentence he is quoting from the framework, and there are three examples in the section

<fsasaki> chris: agree with ashok, it should say s.t. concrete, so that you don't have to look at the examples to figure it out

<fsasaki> asir: I'd like to see a proposal

<fsasaki> ashok: I think david is asking you to fix the wording

<fsasaki> asir: ok, looking at 4.3.3

<fsasaki> .. david wants to get more elaboration about "all distributes over exactlyOne"

<fsasaki> chris: agree

<fsasaki> asir: what david proposes does not help, he wants an explanation what "all distributes over exactlyOne" means

<fsasaki> .. hard to see what satisfies him

<fsasaki> ashok: you should ask him

<fsasaki> asir: will do

<cferris> ACTION: Asir to follow-up with dave Hull to ask what clarification he would like to see [recorded in http://www.w3.org/2007/06/06-ws-policy-minutes.html#action04]

<trackbot> Created ACTION-310 - Follow-up with dave Hull to ask what clarification he would like to see [on Asir Vedamuthu - due 2007-06-13].

[Bug 4584] Clarify how lax mode and ignorable assertions affect the intersection algorithm

Opened by Sergey

Chris walks through the issue

The two policy expressions quoted by Sergey are compatible on lax mode

[Sergey is back]

<cferris> http://www.w3.org/Bugs/Public/show_bug.cgi?id=4584

Sergey: kind of confused .. optional and ignorable assertions mean
... ignorable assertions may be ignored for intersection

<cferris> Provider policy :

<cferris> <Policy>

<cferris> <A/>

<cferris> <B wsp:ignorable="true"/>

<cferris> </Policy>

<cferris> Requester policy :

<cferris> <Policy>

<cferris> <A/>

<cferris> <D wsp:ignorable="true"/>

<cferris> </Policy>

<cferris> Intersected policy using lax mode:

<cferris> <Policy>

<cferris> <A/>

<cferris> <A/>

<cferris> <D wsp:ignorable="true"/>

<cferris> <B wsp:ignorable="true"/>

<cferris> </Policy>

Sergey: points to clarifications from Chris to address 4584

Change 1 is

• Modify Section 3.1, Framework - S/An ignorable policy assertion is an assertion that may be ignored for policy intersection/An ignorable policy assertion is an assertion that may be ignored for purposes of determining the compatibility of alternatives in policy intersection/

Change 2 is

<cferris> http://lists.w3.org/Archives/Public/public-ws-policy/2007May/0289.html

• Add to Section 4.5, Framework – “The behavior implied by an ignorable assertion is expected to be a behavior that need not be engaged for successful interoperation with the entity that includes such ignorable assertions in its policy.”

Sergey: points to some text in the Primer
... plans to start a thread on Primer
... happy with some proposal along those lines

<fsasaki> asir: change 1 is a modification, change 2 and addition

Chris: summarizes the proposed changes to the Framework

Chris is typing change 3

<cferris> S/If two alternatives are compatible, their intersection is an alternative containing all of the assertions in both alternatives/If two alternatives are compatible, their intersection is an alternative containing all of the assertions in both alternatives, regardless of whether or not they are marked with the wsp:Ignorable='true' attribute.

<prasad> This is an issue on the primer but, the resolution is to change the text in framework?

Chris is re-targeting issue 4584 to the Framework

Monica: is it specific to lax mode

<cferris> issue 4584 proposal:

proposed clarification to change 1 is s/policy intersection/policy intersection on lax mode/

<cferris> 1. •

<cferris> Modify Section 3.1, Framework - S/An ignorable policy assertion is an assertion that may be ignored for policy intersection/An ignorable policy assertion is an assertion that may be ignored for purposes of determining the compatibility of alternatives in policy intersection in lax mode/

<monica> c/on/using

<monica> ok in

<cferris> 2. •

<cferris> Add to Section 4.5, Framework – “The behavior implied by an ignorable assertion is expected to be a behavior that need not be engaged for successful interoperation with the entity that includes such ignorable assertions in its policy.”

<cferris> 3. S/If two alternatives are compatible, their intersection is an alternative containing all of the assertions in both alternatives/If two alternatives are compatible, their intersection is an alternative containing all of the assertions in both alternatives, regardless of whether or not they are marked with the wsp:Ignorable='true' attribute.

RESOLUTION: closed issue 4584 with the changes 1-3 above

<scribe> ACTION: Chris to follow with David Hull re issue 4598 [recorded in http://www.w3.org/2007/06/06-ws-policy-minutes.html#action05]

<trackbot> Created ACTION-311 - Follow with David Hull re issue 4598 [on Christopher Ferris - due 2007-06-13].

[onto guidelines issues

[Bug 3988]: Section 8 doesn't illustrate how to design an assertion

<cferris> 3989

[Bug 3989] Suggested Format

<prasad> +1 to closing

<prasad> Don't we have editors actions for all these?

editors have a slew of actions

Resolution: close issue 3989 with the updated Guidelines draft delivered by the editors for the Ottawa F2F

[Bug 3988]: Section 8 doesn't illustrate how to design an assertion

Asir: let's wait until we are done with all the editorial actions and WG level actions against issue 3988 and then take action on this issue

Chris recorded this in issue 3988

[Bug 4566]: Guideline G2 to be reconsidered

<cferris> http://www.w3.org/Bugs/Public/show_bug.cgi?id=4566

<prasad> Description: Guideline G2 [1], slated for Section 5.1 of the Guidelines

<prasad> document [2], reads: “An assertion author should define policy assertions for

<prasad> behaviors that are relevant to compatibility tests, such as web service

<prasad> protocols that manifest on the wire”. It seems to me that this good practice

<prasad> has been overtaken by events around Ignorable (and maybe even the closed/open

<prasad> world discussion of late). As editor, I don’t feel comfortable adding it to the

<prasad> guidelines document without further clarification from the workgroup.

Toufic: one of the best practices says something about .. assertion authors should only design assertions for behaviors that impact the wire

<Ashok> +1 to Toufic

Toufic: may be impacted by open/closed world

<prasad> It is a "should" though, discouraging things w/o wire manifestation

<fsasaki> asir: the BP was not designed before ignorable was introduced

<fsasaki> .. I see no relation to OWA / CWA discussion

<fsasaki> Toufic: saying that you SHOULD do s.t. means you should not do the other thing

<fsasaki> asir: don't follow

<fsasaki> Toufic: I propose a new wording

<scribe> ACTION: Toufic to propose a different wording for Good Practice G2 [recorded in http://www.w3.org/2007/06/06-ws-policy-minutes.html#action06]

<trackbot> Created ACTION-312 - Propose a different wording for Good Practice G2 [on Toufic Boubez - due 2007-06-13].

[Bug 3978]: Clarify if Section 7 on defining new policy attachment mechanisms is necessary

Chris walks through the status

Maryann: made some changes to the Primer re versioning
... has four parts
... versioning of language vs. versioning of assertions

<TRutt__> Fyi from rfc 2119 : SHOULD This word, or the adjective "RECOMMENDED", mean that there

<TRutt__> may exist valid reasons in particular circumstances to ignore a

<TRutt__> particular item, but the full implications must be understood and

<TRutt__> carefully weighed before choosing a different course.

Maryann: has some text for Section 3.10
... may need to revisit this with the current version of the Primer
... want to revisit the proposal and see if there are any outstanding issues to be understood
... diff between versioning the language and assertions aren't clear

<cferris> ACTION: Maryann to review latest Primer and Guildlines with regards to her four points related to [Bug 3978] [recorded in http://www.w3.org/2007/06/06-ws-policy-minutes.html#action07]

<trackbot> Created ACTION-313 - Review latest Primer and Guildlines with regards to her four points related to [Bug 3978] [on Maryann Hondo - due 2007-06-13].

Asir: proposal needs to be updated (cos Guidelines doc has changed significantly) prior to taking any action on issue 3978

Made through the agenda

We made good progress and we have some actions to follow through

Adjourned