W3C

WS-Policy Teleconference 2007-01-03

3 Jan 2007

See also: IRC log

Attendees

Present
Chris_Ferris, whenry_(muted), Maryann_Hondo, Glen, asir, Mark_Temple-Raston, Sergey_Beryozkin, Tom_Rutt, Fabian, Ashok_Malhotra, Frederick_Hirsch, Yakov_Sverdlov, Abbie_Barbir, m2, Prasad_Yendluri, Sanka, Dan_Roth, Umit, Tony_Nadalin
Regrets
Toufic, Mark_Little
Chair
Chris Ferris
Scribe
GlenD

Contents


 

 

<whenry> I'm calling from Skype in Ireland and it's been echoing all day. I've been on vacation.

<scribe> scribeNick: GlenD

<monica> can get now

<monica> couldn't earlier

Review and approval of previous minutes

Minutes are approved.

Future WG meetings

Paul chairs next telcon

F2F in SF - please register if you haven't

Editorial team report

Umit: Docs not ready until next week - two actions outstanding. Others committed and closed.
... Need to check in with Asir - better idea after today's meeting.

Prasad: Should have docs ready...

Chris: Next week ok?

Umit: Yes, next Wed should be no problem.

Review action items

166 - publication primer + guidelines... DONE

<scribe> ACTION: Felix to update the public web pages to point to the refreshed documents. [recorded in http://www.w3.org/2007/01/03-ws-policy-minutes.html#action01]

<trackbot> Created ACTION-177 - Update the public web pages to point to the refreshed documents. [on Felix Sasaki - due 2007-01-10].

ACTION-170 - DONE

ACTION-171 - http://www.w3.org/2005/06/tracker/wspolicy/actions/171 - PENDING

ACTION-172 - http://www.w3.org/2005/06/tracker/wspolicy/actions/172 - PENDING

Chris: Done by next week?

Asir: Possible, yes

ACTION-173 - http://www.w3.org/2005/06/tracker/wspolicy/actions/173 - PENDING

<abbie> good for u

Chris: By EOW

ACTION-174 - http://www.w3.org/2005/06/tracker/wspolicy/actions/174 - PENDING

ACTION-175 - http://www.w3.org/2005/06/tracker/wspolicy/actions/175 - DONE

ACTION-176 - http://www.w3.org/2005/06/tracker/wspolicy/actions/176 - DONE

Last Call Reviews from other WGs

a) C14N 1.1 Last Call review announcement, XML Core WG

Chris: status unknown, carry this over to next wk

Last Call Framework and Attachment Issues

ISSUE 4069: Updating References for Use of xml:id

Monica: This only affects the primer. We added changes that affect 2.8 in the primer - 1) add a reference to xml:id as a third reference mechanism, and 2) add an example of doing so.
... In sec 3.2 (now 3.6) add a chance which acknowledges the use of xml:id.

MS indicates they're fine with this.

<monica> http://lists.w3.org/Archives/Public/public-ws-policy/2006Dec/0081.html

Chris: Any objection to closing 4069 with this proposal?

<silence>

RESOLUTION: Close issue 4069 with the proposal from Monica ( http://lists.w3.org/Archives/Public/public-ws-policy/2006Dec/0081.html)

<scribe> NEW ISSUE 4128: Add References to WSDL 1.1 and WSDL 2.0 Component Syntax, Ashok

http://lists.w3.org/Archives/Public/public-ws-policy/2007Jan/0003.html

<asir> 4069 related editorial action is http://www.w3.org/2005/06/tracker/wspolicyeds/actions/110

Ashok explains the issue referenced in the email.

Chris: Can we just turn this over to the editors?

Ashok: Fine by me.
... Dan mentioned we also need to explain how effective policy is calculated when using these domain expressions.

Chris: Can we split this, just add the refs now and then have another issue for wording on effective policy?

<umit> +1 to splitting this issue

Ashok: Dan and I have already agreed on some words...

Asir: But those were in a different context, should make sure they make sense now.
... Refs appear to be in 3.4.1 - para right b4 second example...

Ashok: Still need at least WSDL 1.1
... Stylistically, would be nice to have the refs at the first use

Chris: Adding reference to the note we're working on seems a done deal. Not closed yet on effective policy wording.

Ashok / Dan meld minds to recall agreement or not

Chris: OK, let's take this offline, and do the editorial changes for the references now.

(Asir adds editorial action for the team to add references earlier)

<asir> related editorial action is http://www.w3.org/2005/06/tracker/wspolicyeds/actions/111

c) NEW ISSUE 4129: Attaching Policies to EPRs, Ashok

http://lists.w3.org/Archives/Public/public-ws-policy/2007Jan/0004.html

<scribe> ACTION: Ashok and Dan to work to come to consensus wording describing effective policy calculation with respect to issue 4128. [recorded in http://www.w3.org/2007/01/03-ws-policy-minutes.html#action02]

<trackbot> Created ACTION-178 - And Dan to work to come to consensus wording describing effective policy calculation with respect to issue 4128. [on Ashok Malhotra - due 2007-01-10].

<umit> that is my recollection as well, Glen.

<cferris> glen: this proposal needs more discussion... at best is not a complete solution

<cferris> ashok: how should we proceed

<cferris> glen: fine with raising the issue though

<asir> F2F discussion pointer - http://www.w3.org/2006/11/09-ws-policy-minutes.html#item09

Glen: The issue was that EPR equivalence becomes an issue when trying to implement the proposed solution. This needs to be at least addressed, if not completely solved.

Asir: (recapitulates F2F minutes)

Glen: Also there's no "<wsa:EndpointReference>" GED
... Propose that Ashok / Glen go off and rewrite this in some way before the group accepts the issue.

<scribe> ACTION: Glen and Ashok to come up with complete wording/proposal for EPR-related LC issue regarding policy attachment. [recorded in http://www.w3.org/2007/01/03-ws-policy-minutes.html#action03]

<trackbot> Created ACTION-179 - And Ashok to come up with complete wording/proposal for EPR-related LC issue regarding policy attachment. [on Glen Daniels - due 2007-01-10].

d) NEW ISSUE 4130: Ignorable assertion must be ignored, Ashok

http://lists.w3.org/Archives/Public/public-ws-policy/2007Jan/0005.html

Ashok: Would like ignorable to be stronger - MustIgnore.
... Sergey seemed to want the stronger version.

<cferris> glen: the 'able' part is intentional... it means you MAY ignore... lax enables this

<cferris> glen: agree not as strong as stating that you MUST ignore, but think that is stupid

<cferris> glen: prefer to leave as is

Umit: +1 to Glen. Had lots of discussion about partitioning clients, and how some want to use the assertion, and others don't even know about it... so we should leave it as is to support this.

<Zakim> cferris, you wanted to take my hat off and make my case

Chris: Speaking as Chris, not as chair.
... Compelling use cases for why you want to expose assertions which don't change wire messages but may want to advertise QoS. These don't impose requirements, but they MAY want to be aware of them.
... Confidentiality example, for instance. Delivery assurance another.
... Consumer may want specifically to do intersection that uses those particular assertions. Yet there may be clients which don't care... so it's nice to have strict/lax to support these cases.

<umit> +1 to the hatless Chris

<FrederickHirsch> +1 to Chris, Glen

Asir: We spent lots of time on this - if we want to reopen, we should have serious grounds to do so....

Ashok: If I as a server publish an ignorable assertion, I don't know whether the client will "really" ignore it or not... would like to be able to say MustIgnore

<Zakim> GlenD, you wanted to mention it's the clients who don't understand these assertions which matter

<cferris> glen: there is a class of assertions that are used for (e.g) configuration... that you don't want the consumer to care or know about... these should be excluded from the publically published policy

<cferris> glen: not use Ignorable

<SergeyB> Chris, minor clarifications : I actually told Ashok that 'lax' mode can be used now to achive the ignorability and I also said wsp: optional would be the only way to achibe ignorability if wsp:ignorable didnt exist

<cferris> glen: if you mark an assertion Ignorable that actually requires some behavior on the part of the policy consumer, then yes, you will have problems

<cferris> glen: if you don't mark an assertion that has no wire manifestation as Ignorable and the consumer doesn't understand the assertion, you don't have intersection and cannot interact

<whenry> +1 to Glen. Use mynamespace:local or mynamespace:MustIgnore

<TRutt_> +1 to Glen

Glen: Ashok, can you give a use-case for MustIgnore? Why would you do that?

Ashok: Legal policy or ad policy... don't want it used as policy selection process.

<monica> let's remember our manners, please

<whenry> And ... ignore it

Glen: But what are they going to do with it?

Ashok: Ignory stuff! Maybe choose not to talk to the service

Glen: Isn't that selection?

Ashok: I can put a pointer to my legal policy in there... Client isn't going to do intersection.

Umit: But it might be a precursor to intersection.

Ashok: I think these things are purely abstract, and you can use them to decide if you want to work with the guy...

Glen: Isn't that the same thing as intersection, at least abstractly?

Ashok: I don't think so.

(discussion continues)

(discussion of whether legal policies will be matched by policy engines or not)

<sanka> PING

PONG

Chris: I should be free to be stupid

<MarkTR> +1 to Chris

<sanka> +1 to Chris from me too ..

Chris: In other words, why should you overly constrain me as to whether I ignore or don't a particular assertion

<monica> again: let's remember our manners, please

Chris: You can "force" people to look at something by NOT marking it as ignorable

Ashok: This is stuff that's "for human eyes only"

Umit: If you have a legal policy which you use to decide whether or not you use a given endpoint, then you are always forced to process the legal policy, yes?

Ashok: Most people don't read EULAS....

Umit: Either you choose to do something with your metadata or not. If you do, then you understand what the QName is for the legal assertion and understand how to use that assertion in determining whether or not you engage with the endpoint.
... This is like being in strict mode

Sergey: Ashok seems to want the user to make the decision on a certain assertion, and not the algorithm to fail before that. We can still acheive this with lax mode.
... Tools can be configured to ask user about unrecognized assertions.

Chris: EULA analogy. Forces you to click before using whatever. As the consumer, I can read it or not before I click.
... That's my choice. This is like strict/lax.

<TRutt_> "for human eyes only" does not match the semantics of ignorable. Perhaps he is asking for a new attribute type "for information only", However I do not see the requirement for such a concept

Glen: EULA isn't right example because you MUST click the button. EULA without a button is a better example of ignorable.
... Can Ashok and I talk about this and come back next week?

Leaving this for next week.

e) NEW ISSUE 4138: Normalization Algorithm is broken, Umit

http://lists.w3.org/Archives/Public/public-ws-policy/2007Jan/0011.html

<Ashok> sounds good, Glen ... i'll send you mail

Umit: Normalization algorithm doesn't handle some cases... example in mail.
... Recommendation is to add one line to the algorithm - when result yields conjunction containing a single assertion or a set, then it's equivalent to a single alternative which contains that result assertion/set.

Fabian: +1 to Umit's issue. Prefer to have algorithm defined in terms of operators.

Asir: Algorithm says to construct "normal form" which is defined elsewhere. Would like more time to consider.

Umit: Normal form isn't theoretically complete.

Asir: There is a sentence there... maybe this is a clarification?

Umit: Shouldn't be part of the recursion... this problem only happens when recursion completes with conjunctive form with no alternatives.

Chris: Everyone understand the issue?
... Please review and consider.

Umit: Not tied, btw, to this particular proposal. Just want to fix the problem.

f) NEW ISSUE 4141: Policy parameter definition is not accurate, Umit

http://lists.w3.org/Archives/Public/public-ws-policy/2007Jan/0013.html

Umit: Framework doc is not accurate about what parameterized assertions are...
... Proposal is to change text to qualify parameters as "things which are not <wsp:Policy>"

Chris: Policy references too, right?

Umit: Cool - friendly amendment is to just watch the namespace.

Dan: Isn't this only for a pre-normalized policy?

Umit: I don't think so

Chris: No Post-Policy-Validation Infoset...

Asir: Data model is close to normal form, and description of parameter is at the data model level...
... So policy references and inclusions have already been processed

<umit> This is very bad, Post schema validation infoset.

Asir: So you could say "normal form" == PPVI

Umit: WHOA! If *we* missed that detail, what about our readers?
... Someone looking at an assertion (XML) should be able to tell what's a parameter and what's nested/referenced....

Asir: We like the proposal, but I think we digressed here.

<cferris> dan: question is whether or not this definition applies to normal form or to the definition of an assertion

<cferris> dan: we may need to talk about the infoset of the assertion in its normal form

<cferris> umit: that is a second level consideration

<cferris> umit: we need to exclude nested policy from the definition of parameters

<maryann_> so can someone propose ammendments?

Chris: Consensus as to the definition of parameter, but do we need to clarify both in term of normal form AND infoset form?

<scribe> ACTION: Umit and Dan to discuss resolution of 4141 and come back with an amended proposal for 2007-01-10. [recorded in http://www.w3.org/2007/01/03-ws-policy-minutes.html#action04]

<trackbot> Created ACTION-180 - And Dan to discuss resolution of 4141 and come back with an amended proposal for 2007-01-10. [on Umit Yalcinalp - due 2007-01-10].

g) NEW ISSUE 4142: Contradictory recommendation for nesting and intersection, Umit

http://lists.w3.org/Archives/Public/public-ws-policy/2007Jan/0014.html

Umit: Recommending an empty <wsp:Policy> in order to help intersection not fail. What does an empty policy really mean?
... seems to be contradictory with 4.3.2...

<cferris> q

Umit: Either last sentence of 4.3.2 is incorrect, or our algorithm for compatibility needs to accommodate an empty policy expression.
... We intended the former - last sentence is wrong.

Maryann: Must still be an alternative which matches, so the reason for this was to supply a valid alternative for matching.

Umit: So you agree with first interpretation?

Maryann: Yes. Should be clearer.

Umit: Might need an example somewhere in guidelines or primer.

Chris: Guidelines/primer issue is secondary
... Intent was clearly not to make things fail based on inclusion/omission of empty policy, we should clean it up.

Asir: +1 there's consensus that last sentence is misleading. Can we just drop it?

Umit: Maryann and I can work on something and then see what the group thinks

<scribe> ACTION: Umit and Maryann to come up with a new version of the wording for the last sentence of sec 4.3.2, as an amended proposal for 4142 [recorded in http://www.w3.org/2007/01/03-ws-policy-minutes.html#action05]

<trackbot> Created ACTION-181 - And Maryann to come up with a new version of the wording for the last sentence of sec 4.3.2, as an amended proposal for 4142 [on Umit Yalcinalp - due 2007-01-10].

<scribe> ACTION: Umit to file a new issue against primer/guidelines suggesting the need for examples of intersection with/without empty wsp:Policy (see issue 4142) [recorded in http://www.w3.org/2007/01/03-ws-policy-minutes.html#action06]

<trackbot> Created ACTION-182 - File a new issue against primer/guidelines suggesting the need for examples of intersection with/without empty wsp:Policy (see issue 4142) [on Umit Yalcinalp - due 2007-01-10].

8. Primer Doc Issues

a) ACTION-152 Review 4.4.8 with respect to the addition of the ignorable attribute ("treasure island") due December 1, David Orchard

<scribe> DONE. See:

http://lists.w3.org/Archives/Public/public-ws-policy/2006Dec/0070.html

Status: David outlined his proposal at the Dec 20 meeting. The WG has until Jan 3 to decide if it should be adopted.

Chris: Can we review this by next week?

Defer discussion of ACTION-152 until next week.

b) (NEW) ISSUE 4041: Update primer to mention ignorable as needed, Frederick

http://www.w3.org/Bugs/Public/show_bug.cgi?id=4041

http://lists.w3.org/Archives/Public/public-ws-policy/2006Dec/0026.html

<scribe> PENDING

c) [NEW ISSUE] 4103 Questionable use of Contoso Ltd in Primer, Chris Ferris

http://www.w3.org/Bugs/Public/show_bug.cgi?id=4103

http://lists.w3.org/Archives/Public/public-ws-policy/2006Dec/0077.html

See thread ending at:

http://lists.w3.org/Archives/Public/public-ws-policy/2006Dec/0102.html

Chris: Contoso is registered TM of Microsoft?

Asir: Can check on this.

Chris: Had suggested example.com, using what a lot of other people use.
... Example.com as a company name is not registered, though.

<scribe> ACTION: Asir to research ownership of "Contoso" name/trademark. [recorded in http://www.w3.org/2007/01/03-ws-policy-minutes.html#action07]

<trackbot> Created ACTION-183 - Research ownership of \"Contoso\" name/trademark. [on Asir Vedamuthu - due 2007-01-10].

Chris: Should have explicit permission, at least, if we're going to use a non-W3C-registered name.

<umit> why can't we use Company A as we do in the Guidelines?

<cferris> ACTION: Chris to follow-up on arch of www doc's use of a company name [recorded in http://www.w3.org/2007/01/03-ws-policy-minutes.html#action08]

<trackbot> Created ACTION-184 - Follow-up on arch of www doc\'s use of a company name [on Christopher Ferris - due 2007-01-10].

<Nadalin> can't use that !

<Nadalin> I have it that registerd

<maryann_> aw come on

<maryann_> :-)

<Nadalin> NoNameCompany

Chris: Next week we'll do guidelines issues and proposals for new issues.

ADJOURN

<asir> http://www.nonamecompany.com/ is taken

Summary of Action Items

[NEW] ACTION: Ashok and Dan to work to come to consensus wording describing effective policy calculation with respect to issue 4128. [recorded in http://www.w3.org/2007/01/03-ws-policy-minutes.html#action02]
[NEW] ACTION: Asir to research ownership of "Contoso" name/trademark. [recorded in http://www.w3.org/2007/01/03-ws-policy-minutes.html#action07]
[NEW] ACTION: Chris to follow-up on arch of www doc's use of a company name [recorded in http://www.w3.org/2007/01/03-ws-policy-minutes.html#action08]
[NEW] ACTION: Felix to update the public web pages to point to the refreshed documents. [recorded in http://www.w3.org/2007/01/03-ws-policy-minutes.html#action01]
[NEW] ACTION: Glen and Ashok to come up with complete wording/proposal for EPR-related LC issue regarding policy attachment. [recorded in http://www.w3.org/2007/01/03-ws-policy-minutes.html#action03]
[NEW] ACTION: Umit and Dan to discuss resolution of 4141 and come back with an amended proposal for 2007-01-10. [recorded in http://www.w3.org/2007/01/03-ws-policy-minutes.html#action04]
[NEW] ACTION: Umit and Maryann to come up with a new version of the wording for the last sentence of sec 4.3.2, as an amended proposal for 4142 [recorded in http://www.w3.org/2007/01/03-ws-policy-minutes.html#action05]
[NEW] ACTION: Umit to file a new issue against primer/guidelines suggesting the need for examples of intersection with/without empty wsp:Policy (see issue 4142) [recorded in http://www.w3.org/2007/01/03-ws-policy-minutes.html#action06]
 
[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.127 (CVS log)
$Date: 2007/01/10 17:11:14 $