RIF Telecon 5-Aug-08 -- 05 Aug 2008

<LeoraMorgenstern> I'll volunteer.

<ChrisW> Scribe: AdrianP

Admin

<ChrisW> http://lists.w3.org/Archives/Public/public-rif-wg/2008Jul/att-0217/29-July-2008-rif-minutes.html

ChrisW: Objections to accepting minutes?

<ChrisW> PROPOSED: Accept minutes of 29-July telecon

<ChrisW> RESOLVED: Accept minutes of 29-July telecon

Liason

ChrisW: Status of SBVR?

Christian: SBVR currently an adopted standard
... Large document, 400 pages, it is computer independent level specification of business rules
... RIF is at the execution level, SBVR is not
... Framework for analysing the business semantics of rules

<Harold> http://www.omg.org/spec/SBVR/1.0/

Harold: use common logic as underlying logic as far as I know

Christian: there is a mapping at least part of it

<Harold> http://en.wikipedia.org/wiki/Semantics_of_Business_Vocabulary_and_Business_Rules

ChrisW: Translation between SBVR and RIF?

<Harold> "Common Logic ? draft standard 24707, currently being fast-tracked by ISO."

<Harold> "SBVR is aligned with Common Logic ? draft standard 24707, currently being fast-tracked by ISO."

Christian: If you could map SBVR to executable rules it can be mapped to RIF

Adrian: HCLS has recharted and will start again in September

Axel: will talk about rif:text next week with the OWL group

Action Review

Michael: Send an eMail to the list about FLD semantics for PRD

F2F11

ChrisW: most prefer to have meeting in Manhattan

<AxelPolleres> Are there any Hotel recommendations?

ChrisW: not far from Central Park and Broadway
... Will put information on the webpage this week

Core

Christian: Core is one of the documents on the list. What do we do with it? Is it is on priority?

ChrisW: DTB, PRD, FLD, BLD
... TestCases and Core just some initial work

Christian: We had a discussion in July about Core
... concensus was that we could need a Core document if we have a use case for Core
... Something smaller than the intersection of PRD and BLD

ChrisW: Is the intersection clear?

Christian: Not completely yet. But can be figured out from the current working drafts

Dave: Core something minimum Horn Clause, no equality, ..

<AxelPolleres> no function symbols?

Christian: We would need a seperate document, we would need a use case for Core
... if there is no use case for core we might question the need for a seperate document

Dave: right
... We would be interested into interchange from Jena, simple Jena rules

ChrisW: What aspects of BLD would you drop?

Dave: Core question is function symbols or not
... recursion

<csma> Dave proposed interchange between simple Jena rules and N3 as a use case for Core

Dave: ability to call built-ins, but not unrestricted functions and data structures

ChrisW: Other features which should be droped?

Christian: You mentioned membership and classification

Dave: Depends on the goa of Core
... If minimum intersection is the goal it should be removed

Christian: You can not remove the ability to call built-ins from PRD
... You can not remove frames, named arguments from PRD
... to be compatible with Clips using named arguments

ChrisW: Frames seem to be in the intersection

Dave: right

Christian: PRD in the first WD has named arguments since they are in BLD

ChrisW: list, frames should be in Core?

Harold: Yes, I think so

Dave: Have to think about it, if I can be editor

Harold: I would be also interested

Jos: Removing atomic formulas like classification, member, equality is good idea
... Remove also disjunction from the body of rules
... Makes processing easier

<AxelPolleres> +1 to keep out disjunction as well

<Harold> I totally agree with removing disjunction from bodies of rules in Core.

Test Cases

<Harold> Adrian, I could scribe temporarily

ChrisW: Test Cases are important since we are a standardization body
... We need a test cases document
... We postponed it to finish last call

<csma> http://www.w3.org/2005/rules/wiki/Category:Test_Case

<Harold> Adrian: Two wiki pages: old and new wiki pages

<csma> http://www.w3.org/2005/rules/wg/wiki/Arch/Test_Cases

<csma> ACTION: ChristopherW to add a link to test cases pages on the RIF front page [recorded in http://www.w3.org/2008/08/05-rif-minutes.html#action01]

<trackbot> Sorry, couldn't find user - ChristopherW

<Harold> Chris: Links from main pages should be added.

<csma> ACTION: chris to add a link to test cases pages on the RIF front page [recorded in http://www.w3.org/2008/08/05-rif-minutes.html#action02]

<trackbot> Created ACTION-547 - Add a link to test cases pages on the RIF front page [on Christopher Welty - due 2008-08-12].

<Harold> Adrian: Added interface for test cases.

http://www.w3.org/2005/rules/wiki/Category:Test_Case

<Harold> ... Concrete implementation is platform-specific.

<Harold> ... But general interface is independent.

http://www.w3.org/2005/rules/wg/wiki/Arch/Test_Cases

<Harold> Adrian: Ex. in section "Categories of RIF Dialect test cases" at above URL.

<Harold> ... Can test if all test cases succeed.

<AxelPolleres> Do you have something like SPROT in mind? i.e. we'd need to define an interface, result format, etc.

<Harold> ... E.g. built-in tests are done in each rule engine, reached from the interface.

<StellaMitchell> sorry, wrong button

<StellaMitchell> calling back

<Harold> Chris: Time and bandwidth? Both test cases editor and PRD editor?

<AxelPolleres> cf. http://www.w3.org/TR/rdf-sparql-protocol/

<Harold> Adrian: Depends a bit on kinds of test cases.

<Harold> ... E.g., there are very general test cases.

<Harold> Chris: Contingent on what we want to do with test cases document.

<Harold> Adrian: Yes, depends a bit on how general these test cases are.

<Harold> Stella: Agree with Adrian.

<Harold> Chris: Would you contribute to doc?

<Harold> Stella: Yes, but cannot pile up all these test cases.

<Harold> Chris: Of course, collaborative.

<Harold> Stella: We should test both RIF consumers and producers.

<Harold> Chris: Good point.

<Harold> Gary: Do we have any (new) test cases?

<Harold> ... interesting ones?

<Harold> ... Corner cases of BLD and PRD.

<Harold> Adrian: The idea is that only the interface is given, from which specific ones are generated.

<Harold> Chris: Tricky to do test cases for an interchange format.

<Harold> Adrian: Idea like for Built-ins. Library + Specific.

<Harold> ... Could also contribute to use case doc with Stella. But first place in my priorities is PRD.

<Harold> Chris: OK

<Harold> Gary: E.g. you could write ground facts and rules and test these.

here are some examples http://www.w3.org/2005/rules/wiki/Category:Test_Case

<Harold> Chris: Not so much translation test cases.

<Harold> Gary: Right. Need entailment test to help with understanding spec.

<AxelPolleres> By entailment tests we mean just *qround* queries? but no non-ground queries?

<Harold> Christian: We need to make sure that what we specify works.

<Harold> ... need exanples.

<Harold> Gary: Yes, good test cases can show model-theoretic semantics.

<Harold> ... another way to show the implementation.

<Harold> Christian: We should have test cases but they cannot say how things must be.

<Harold> ... You have to say something about semantics.

<Harold> Chris: No one is for translation test cases?

<Harold> Christian: Well, we need them but it's tricky: Several ways to translate some rules.

<Harold> Chris: Just want to clarify if test cases are for exemplifying entailments or translations.

<Harold> Christian: Mostly translation. Semantics must be linked to translation, too.

<LeoraMorgenstern> If Stella is interested in help, I could help out.

<Harold> Chris: Any one else interested become a test cases co-editor? Please let chairs know.

<StellaMitchell> yes

<StellaMitchell> yes Leora

<LeoraMorgenstern> Okay, great.

<Harold> Adrian: Will move the wiki page as discussed.

<ChrisW> ACTION: stella to work on test cases [recorded in http://www.w3.org/2008/08/05-rif-minutes.html#action03]

<trackbot> Created ACTION-548 - Work on test cases [on Stella Mitchell - due 2008-08-12].

<ChrisW> ACTION: adrian to work on test cases [recorded in http://www.w3.org/2008/08/05-rif-minutes.html#action04]

<trackbot> Sorry, amibiguous username (more than one match) - adrian

<trackbot> Try using a different identifier, such as family name or username (eg. agiurca, apaschke)

<ChrisW> ACTION: apaschke to work on test cases [recorded in http://www.w3.org/2008/08/05-rif-minutes.html#action05]

<trackbot> Created ACTION-549 - Work on test cases [on Adrian Paschke - due 2008-08-12].

FLD

Michael: FLD needs some improvement in the ways it is defined
... right now it is not extensible enough
... e.g. connectives can not be easily added
... define signatures for connectives
... treat connectives like symbols

ChrisW: should we start trying to get people from other communities to look at FLD

Michael: yes, would be good

<AxelPolleres> I could contact ALP (Association of Logic Programming)

Michael: I don't know how lively this publication has been circulated
... Sandro's announcement has not been widely circulated

ChrisW: advertising and circulation is our job. We should send it to the lists

Michael: also BLD

<Harold> Sandro had sent this http://lists.w3.org/Archives/Public/semantic-web/2008Aug/0002.html

ChrisW: send an announcement to the mailing where you are on

Christian: we must make sure that any interested group gets informed about last call
... If you know any group who might be interested, please inform them

ChrisW: you might use the text from Sandro's mail
... Let's compile a list of experts

Christian: We should start a page so that people and mailing lists are not contacted twice

<ChrisW> ACTION: christian to start a wiki page for recording mailing list announcements sent [recorded in http://www.w3.org/2008/08/05-rif-minutes.html#action06]

<trackbot> Created ACTION-550 - Start a wiki page for recording mailing list announcements sent [on Christian de Sainte Marie - due 2008-08-12].

DTB

ChrisW: Casting function described in DTB from IRI to String is not a function
... We discussed it some weeks ago; no strong oppinion

Axel: Preferd to leave it as it is defined in DTB now

ChrisW: 3 people prefered to remove it; drafted a resolution

<ChrisW> PROPOSED: Remove section 4.3.5 (casting function for rif:iri) from DTB.

Axel: I wanted to have it published in the working draft and see if there is any feedback

<Harold> I agree.

Axel: I would not drop it immediately and wait for feedback

<AxelPolleres> :-)

ChrisW: alternative would be to drop it and then wait for feedback
... We can later reopen it

<ChrisW> PROPOSED: Remove section 4.3.5 (casting function for rif:iri) from DTB.

<josb> +1

<AxelPolleres> -\epsilon

<Harold> 0

<DaveReynolds> 0

<ChrisW> +1

<GaryHallmark> 0

<MichaelKifer> +1

<StellaMitchell> +1

<ChrisW> CSMA: +1

<ChrisW> RESOLVED: Remove section 4.3.5 (casting function for rif:iri) from DTB.

<ChrisW> ACTION: axel to remove section 4.3.5 from dtb [recorded in http://www.w3.org/2008/08/05-rif-minutes.html#action07]

<trackbot> Created ACTION-551 - Remove section 4.3.5 from dtb [on Axel Polleres - due 2008-08-12].

Axel: Editor notes need to be done
... I would summarize then and categorize them in an eMail

ChrisW: Guard predicates need to be discussed as far as I remember
... Yes, please summarize them

<ChrisW> ACTION: axel to summarize open DTB issues in editors notes [recorded in http://www.w3.org/2008/08/05-rif-minutes.html#action08]

<trackbot> Created ACTION-552 - Summarize open DTB issues in editors notes [on Axel Polleres - due 2008-08-12].

PRD

Christian: lots of issues, editor notes
... we started to discuss one of them
... still many things to do till last call
... all issues need to be discussed and we need to find a consensus
... I will be on vacations by the end of August; other ILog staff members
... so no input from the ILog side

ChrisW: Gary and Adrian have voluntiered to help as editors

Christian: DTB should work for PRD as well

Michael: Semantics for conditions is the same for PRD and BLD
... rephrase the semantics of PRD according to BLD

ChrisW: update semantics of condition part of PRD

<ChrisW> ACTION: adrian to work on making PRD conditions equivalent to BLD conditions [recorded in http://www.w3.org/2008/08/05-rif-minutes.html#action09]

<trackbot> Sorry, amibiguous username (more than one match) - adrian

<trackbot> Try using a different identifier, such as family name or username (eg. agiurca, apaschke)

<ChrisW> ACTION: apaschke to work on making PRD conditions equivalent to BLD conditions [recorded in http://www.w3.org/2008/08/05-rif-minutes.html#action10]

<trackbot> Created ACTION-553 - Work on making PRD conditions equivalent to BLD conditions [on Adrian Paschke - due 2008-08-12].

Adrian: What about presentation syntax? Open questions use :- or =>

Christian: Wait for feedback

ChrisW: Shall we lively ask for feedback?

Christian: Get as many and as much early feedback as we can

ChrisW: People should start to disseminate to groups, companies, people

Christian: On the actions side we need a strawman about expressive actions
... Gary?

Gary: Skolem functions could be used

ChrisW: action about object creation action

<ChrisW> ACTION: gary to work on object creation action for PRD [recorded in http://www.w3.org/2008/08/05-rif-minutes.html#action11]

<trackbot> Created ACTION-554 - Work on object creation action for PRD [on Gary Hallmark - due 2008-08-12].

Gary: interoperate and exchange frame axioms

Christian: I would be concerened, since we would define RIF in RIF

<ChrisW> ACTION: gary to draft email on "frame axioms" for capturing semantics of PRD frames [recorded in http://www.w3.org/2008/08/05-rif-minutes.html#action12]

<trackbot> Created ACTION-555 - Draft email on \"frame axioms\" for capturing semantics of PRD frames [on Gary Hallmark - due 2008-08-12].

- DRAFT -

RIF Telecon 5-Aug-08

05 Aug 2008

Attendees

Contents