RIF Telecon 11-Nov-08 -- 11 Nov 2008

<ChrisW> Scribe: Harold

<ChrisW> http://lists.w3.org/Archives/Public/public-rif-wg/2008Nov/att-0008/2008-11-04-rif-minutes.html

<ChrisW> PROPOSED: Approve last weeks minutes

<ChrisW> RESOLVED: Approve last weeks minutes

Admin

<ChrisW> ACTION: Chris to change times to standard time in agenda [recorded in http://www.w3.org/2008/11/11-rif-minutes.html#action01]

<trackbot> Created ACTION-643 - Change times to standard time in agenda [on Christopher Welty - due 2008-11-18].

Liason

Axel: Mapping rules written in a RIF dialect would be fine for his liaison.

Public comments

<sandro> Sandro: Jos is going to visit OWL-WG to talk about ID/IDREF.

Chris: Three outstanding responses.
... Just point OK to the changed BLD text.

<ChrisW> ACTION: chris to draft comment to ok [recorded in http://www.w3.org/2008/11/11-rif-minutes.html#action02]

<trackbot> Created ACTION-644 - Draft comment to ok [on Christopher Welty - due 2008-11-18].

<ChrisW> ACTION: Chris to send WL response [recorded in http://www.w3.org/2008/11/11-rif-minutes.html#action03]

<trackbot> Created ACTION-645 - Send WL response [on Christopher Welty - due 2008-11-18].

<sandro> (I'm having trouble with international dialing. Anyone else have another phone line they can use to call Christian? msg me if you need the number.)

<ChrisW> http://www.w3.org/2005/rules/wiki/Response_to_RAK

<ChrisW> ACTION: Gary to draft response to http://www.w3.org/2005/rules/wiki/Response_to_RAK [recorded in http://www.w3.org/2008/11/11-rif-minutes.html#action04]

<trackbot> Created ACTION-646 - Draft response to http://www.w3.org/2005/rules/wiki/Response_to_RAK [on Gary Hallmark - due 2008-11-18].

Action review

<LeoraMorgenstern> 614; I'm just sending out hte email

Michael, should we put changes always into section 2 of specs? _ thought it would always be the last appendix: http://www.w3.org/2005/rules/wiki/BLD#Changes_Since_the_Last_Call_Version_.282008-07-30.29

<Stella> should manifest be higher priority than that translator?

<Stella> I mean, shouldn't

rdf:text

Axel: Found agreement with OWL WG on rdf:text.

<ChrisW> ACTION: axel to add editors note regarding status of finite/infinite characters [recorded in http://www.w3.org/2008/11/11-rif-minutes.html#action05]

<trackbot> Created ACTION-647 - Add editors note regarding status of finite/infinite characters [on Axel Polleres - due 2008-11-18].

<sandro> PROPOSED: Publish rdf:text as a FPWD, jointly with OWL-WG, given the Editor's Note about infinite character sets gets added.

<sandro> +1 (W3C)

<josb> http://www.w3.org/2007/OWL/wiki/InternationalizedStringSpec

<AxelPolleres> http://www.w3.org/2007/OWL/wiki/InternationalizedStringSpec

<sandro> PROPOSED: Publish rdf:text as a FPWD, jointly with OWL-WG, given the Editor's Notes about infinite character sets (per Chris) and about compat with XML schema (per Jos) gets added.

<ChrisW> ACTION: axel to add editors note on xml schema compatibility [recorded in http://www.w3.org/2008/11/11-rif-minutes.html#action06]

<trackbot> Created ACTION-648 - Add editors note on xml schema compatibility [on Axel Polleres - due 2008-11-18].

<sandro> Jos: The infiniteness of the character set, as well as the compatibilty with xml string -- whether the string spart of the lex & val space should be the same as xs:string -- are under discussion.

Axel: Sandro: Best as two separate Ed's notes.

Jos/Sandro: One should be fine if both aspects are kept together.

<sandro> PROPOSED: Publish rdf:text as a FPWD, jointly with OWL-WG, adding an editor's note about infinity (actions Jos, Axel)

<sandro> +1

<sandro> +1 (W3C)

<AxelPolleres> +1 (DERI)

+1 (NRC)

<josb> +1 (FUB)

<LeoraMorgenstern> +1

<Gary> +1 (Oracle)

<LeoraMorgenstern> +1 (IBM)

<StuartTaylor> +1 (Aberdeen)

<mkifer> +1

<sandro> RESOLVED: Publish rdf:text as a FPWD, jointly with OWL-WG, adding an editor's note about infinity (actions Jos, Axel)

<ChrisW> ACTION: send proposed editors note to rdf:text [recorded in http://www.w3.org/2008/11/11-rif-minutes.html#action07]

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

Jos: Will send editor's note, Axel should track it.

<ChrisW> ACTION: debruij2 send proposed editors note to rdf:text [recorded in http://www.w3.org/2008/11/11-rif-minutes.html#action08]

<trackbot> Sorry, couldn't find user - debruij2

<ChrisW> ACTION: jdebruij2 send proposed editors note to rdf:text [recorded in http://www.w3.org/2008/11/11-rif-minutes.html#action09]

<trackbot> Created ACTION-649 - Send proposed editors note to rdf:text [on Jos de Bruijn - due 2008-11-18].

<ChrisW> ACTION: Sandro to publish rdf:text [recorded in http://www.w3.org/2008/11/11-rif-minutes.html#action10]

<trackbot> Created ACTION-650 - Publish rdf:text [on Sandro Hawke - due 2008-11-18].

PRD

Chris: Trying to call Christian's cell...

Gary: Most actions from Orlando were put into the PRD document.
... E.g., our understanding of conflict resolution strategies is now in the text.
... People could start draft on wiki and give feedback.

<ChrisW> no response from csma's phone

Sandro: Does it look like easily implementable on exist. PR engines?

Gary: Some aspects are not trivial to implement.
... E.g., tagging Obj. with particular class via Member is not usual in PR engines.
... Better than removing Member from Core.

<ChrisW> *PROPOSED:* In RIF-PRD, the conflict resolution strategy for a set of rules will be indicated using keywords, attached to the top-level group. RIF-PRD 1.0 will specify only one normative conflict resolution strategy, as specified in csma's email [9] (essentially: refraction+priority+recency). RIF-PRD 1.0 MAY specify other conflict resolution strategies (and corresponding keywords) for suggested use, but these will not be conformance points. This resolution and t

Sandro: Sent email about two issues with this PROPOSED.
... 1) The term 'keyword' is not great.

<sandro> Sandro: I just want it clarified that the strategy could be indicated by being part of the element name

<josb> I sent the email to public-rdf-text

Sandro: In XML it seems to imply it would become an attribute, while it may well become a subelement.

<sandro> Gary: I don't think we're saying anything one way or the other about that.

<ChrisW> *PROPOSED:* In RIF-PRD, the conflict resolution strategy for a set of rules will be indicated in some way attached to the top-level group. RIF-PRD 1.0 will specify only one normative conflict resolution strategy, as specified in csma's email [9] (essentially: refraction+priority+recency). RIF-PRD 1.0 MAY specify other conflict resolution strategies (and corresponding keywords) for suggested use, but these will not be conformance points. This resolution and the c

<ChrisW> PROPOSED: In RIF-PRD, the conflict resolution strategy for a set of rules will be indicated in some way attached to the top-level group. RIF-PRD 1.0 will specify only one normative conflict resolution strategy, as specified in csma's email [9] (essentially: refraction+priority+recency). RIF-PRD 1.0 MAY specify other conflict resolution strategies (and corresponding keywords) for suggested use, but these will not be mandatory. This resolution and the complementar

<ChrisW> PROPOSED: In RIF-PRD, the conflict resolution strategy for a set of rules will be indicated using keywords, attached to the top-level group. RIF-PRD 1.0 will specify only one normative conflict resolution strategy, as specified in csma's email [9] (essentially: refraction+priority+recency). RIF-PRD 1.0 MAY specify other conflict resolution strategies (and corresponding keywords) for suggested use, but these will not be conformance points.

<ChrisW> PROPOSED: In RIF-PRD, the conflict resolution strategy for a set of rules will be indicated in some way, associated with the top-level group. RIF-PRD 1.0 will specify only one normative conflict resolution strategy, as specified in csma's email [9] (essentially: refraction+priority+recency).

<sandro> Sandro: So this means you can't merge rulesets.... but that's the state of the art anyway.

<ChrisW> PROPOSED: RIF-PRD 1.0 MAY specify other conflict resolution strategies for suggested use, but these will not be mandatory.

<ChrisW> PROPOSED: Close Issue-64

<sandro> issue-64

<sandro> issue-64?

<trackbot> ISSUE-64 -- Conflict resolution strategies to be covered by PRD? -- OPEN

<trackbot> http://www.w3.org/2005/rules/wg/track/issues/64

<sandro> +0

<Gary> +1

<sandro> +0 on all three at once

<AxelPolleres> 0

<ChrisW> +1

<ChrisW> RESOLVED: In RIF-PRD, the conflict resolution strategy for a set of rules will be indicated in some way, associated with the top-level group. RIF-PRD 1.0 will specify only one normative conflict resolution strategy, as specified in csma's email [9] (essentially: refraction+priority+recency).

<ChrisW> RESOLVED: RIF-PRD 1.0 MAY specify other conflict resolution strategies for suggested use, but these will not be mandatory.

<ChrisW> RESOLVED: Close Issue-64

<ChrisW> ACTION: chris to close issue 64 [recorded in http://www.w3.org/2008/11/11-rif-minutes.html#action11]

DTB

<trackbot> Created ACTION-651 - Close issue 64 [on Christopher Welty - due 2008-11-18].

Axel: Let's go thru the Ed's notes.

Basically Remove: We might introduce additional shortcuts, e.g. for rif:text

Remove: rif:text (in particular, its identifying IRI) is an AT RISK feature.

Re: The naming convention for guard predicates, ...

Axel: Would just leave text unchanged and remove Ed's note.
... When you define your own dialect, should you be encouraged to reuse pred/func prefixes.
... Who is authorized to reuse such namespace prefixes?

Sandro: W3C.

Axel: Should we say new W3C specs should use these namespace prefixes?

Chris: Give advice. Use word "should" or "should not".

<AxelPolleres> W3C specs MAY reuse, 3rd party dialects MUST NOT reuse the pred and func namespace

<AxelPolleres> ?

SHOULD and SHOULD NOT?

Re: It was noted in discussions of the working group, that except guard ... SPARQL's datatype function ... needed.

Axel: Which value to return?
... Can not emulate the same behavior.
... Even with a predicate.
... SPARQL's datatype functions cannot be emulated.

<josb> that's because SPARQL's def is broken

Chris: But what about adding a predicate (not a function), which can bind all these multiple values?
... Was proposed by Dave Reynolds.
... Single predicate for checking types instead of one for each.

Re: The naming convention for negative guard predicates ...

Axel: Same thing.

Re: In the following, we adapt several cast functions from [XPath-Functions].

Axel: Only referring to conversion table.
... Would need more refined explanation.
... Will try to resolve and have it reviewed.

Re: We might split this subsection into separate subsections per casting function...

Axel: TBD

Re: The cast from rif:text to xs:string is still under discussion, ...

Chris: This fct stays here.

Axel: OK.

Re: Conversion from rif:iri to xs:string and vice versa is still under discussion ...

Axel: If everyone is fine as is, we can leave it.
... Strange thing is that can be viewed in both conversion directions.
... Only one is intended.
... Use case is creation IRI from string.

Sandro: Comes also up in POWDER.
... They have various such conversion needs.

Axel: Only the predicate NAME is a problem.
... It's fine to use the same predicate.

Sandro: Think has just been called IRI.
... Gives you IRI of anything.
... Need a bunch of test cases.
... Tricky 'corner' of the language.

Axel: Could write some, but Jos is better in this.
... OK to rename it into iristring?

Jos/Sandro: Yes.

Chris: Potential unexpected results?

Jos: Only some IRIs you dont expect from strings, and vice versa.
... Will do.

Sandro: Some should use RDF.

Jos/Axel: OK.

<ChrisW> ACTION: jos to write some more pathological test cases for potentially confusing uses of iri-string predicate [recorded in http://www.w3.org/2008/11/11-rif-minutes.html#action12]

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

<trackbot> Try using a different identifier, such as family name or username (eg. jdebruij2, jderoo)

<ChrisW> ACTION: jdebruij2 to write some more pathological test cases for potentially confusing uses of iri-string predicate [recorded in http://www.w3.org/2008/11/11-rif-minutes.html#action13]

<trackbot> Created ACTION-652 - Write some more pathological test cases for potentially confusing uses of iri-string predicate [on Jos de Bruijn - due 2008-11-18].

Re: The following treatment of built-ins which may have multiple arities ...

Axel: Would propose to keep multiple arities.

Harold: We need this for min and max functions (for uncertainty encoding etc.).

Chris: Problem: Would you need infinitely many functions, then, semantically?

Axel: No.

Chris: Seems fine as a syntactic convention.

Axel: If we allow variable arities in External, then we are done, too.

<ChrisW> ACTION: chris to open issue on polyadic functions per not in DTB 3.5 [recorded in http://www.w3.org/2008/11/11-rif-minutes.html#action14]

<trackbot> Created ACTION-653 - Open issue on polyadic functions per not in DTB 3.5 [on Christopher Welty - due 2008-11-18].

<ChrisW> ACTION: Chris to put remaining DTB editors notes on agenda [recorded in http://www.w3.org/2008/11/11-rif-minutes.html#action15]

Chris: Remaining four Ed's notes in DTB moved to next telecon.

PICK SCRIBE

Axel: Add link to issue.

<ChrisW> scribe for next week: leora

<trackbot> Created ACTION-654 - Put remaining DTB editors notes on agenda [on Christopher Welty - due 2008-11-18].

- DRAFT -

RIF Telecon 11-Nov-08

11 Nov 2008

Attendees

Contents