Admin

<ChrisW> Scribe: MichaelKifer

<ChrisW> http://lists.w3.org/Archives/Public/public-rif-wg/2008Aug/att-0107/19082008-rif-minutes.html

<ChrisW> PROPOSED: Accept minutes of Aug 19 telecon

<ChrisW> RESOLVED: Accept minutes of Aug 19 telecon

Liason

<ChrisW> Scribe: Harold

Sandro: OWL-RL is a profile for OWL coupled with rule engines.
... It may fit BLD or FLD.

<DaveReynolds> I would hope OWL-RL fits within Core

Chris: Can we ask them to publish such rule bases in RIF XML syntax?

Sandro: Yes. at some point.

Mike: Is it renamed from earlier name?

Sandro: Evocative of RuLes.

<JeffP> QL

<AxelPolleres> URL?

<JeffP> http://www.w3.org/2007/OWL/wiki/Profiles

<sandro> yeah, more specifically: http://www.w3.org/2007/OWL/wiki/Profiles#OWL_2_RL

Action review

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

<AxelPolleres> ... need to check the implications of Botris Motik's reply to me.

<ChrisW> http://www.w3.org/2005/rules/wiki/Response_to_PPS4

Chris: Axel keep looking at comments wrt DTB relationships.

<ChrisW> ACTION: axel to look at DTB comments in Peters message [recorded in http://www.w3.org/2008/08/26-rif-minutes.html#action01]

<trackbot> Created ACTION-568 - Look at DTB comments in Peters message [on Axel Polleres - due 2008-09-02].

<AdrianP> ok, thanks Sandro

<AdrianP> http://www.w3.org/2005/rules/wiki/PRD#Rules_instantiation:_INSTANTIATE

Adrian: Have reformulated, interacting with Christian, the semantics of the PRD instantiate function.

F2F11

Harold: Who is in which hotel?

<sandro> Holiday Inn Midtown at $306/night

Mike: NJ, $169

<sandro> Holiday Inn Midtown at $306/night 440 WEST 57TH ST 1-212-5818100

<AxelPolleres> Portland Square Hotel (47th street), sharing with jos... makes it cheaper.

Test Cases

Leora: Issue of more user friendly PS.
... Adrian has proposed improvements.
... We want to bring up this issue.

Chris: Current PS makes it difficult to talk about RIF.

<csma> Hear, hear!

Leora: It's plain hard. Extra burdon on users.
... Would be good to have other syntax.

<ChrisW> http://www.w3.org/2005/rules/wiki/UCR#Use_Cases

Leora: Maybe with preprocessor to map to current PS.

Chris: What you use is a bunch of shortcuts with "if...do..." production rules.

<AxelPolleres> I had some comments on the shortcuts suggested for arithmetics and comparison... http://lists.w3.org/Archives/Public/public-rif-wg/2008Aug/0124.html not sure whether the mail was clear, but I can explain it.

Adrian: Currently Prolog symbol for "is implied by".

<AxelPolleres> What about triple syntax a la N3?

Chris: Also "->" is not a good symbol for slots.

Axel: Some examples use arithm shortcuts.

<GaryHallmark> what about colon instead of ->, e.g. ?o[k1:v1 k2:v2]

<AdrianP> you mean, e.g. ">" "<" "!=" "=" "<=" ">=" as in programming languages

<AdrianP> instead of e.g. External(pred:numeric-less-than(1 2))

<csma> Why not =, since we currently have only an equality accessor (for slots)?

<csma> as in object[property=value]?

Adrian: Parser can figure out distinctions via arg types (e.g. two numbers).

Axel: Makes sense.

Sandro: I'm highly dubious.
... It would be a half-way rule language.
... Lets either use XML or a real rule language.
... Run UC: Only XML.
... Read UC: Use your own favorite syntax

<AdrianP> the main intention of the abridged syntax is to make it easy to write down and read rules

<AxelPolleres> in xpath, etc you have '+','-', but type-checking of the operators is supposed to take place before it is chosen which add or subtract function is actually meant.

Chris: Requires a tool.

Sandro: Yes. Need bootstrapping.

Christian: Two questions:

<AxelPolleres> ... same for comparisons, i.e., any xpath processor should be able to deal with +,-, >, <, etc. for various datatypes.

Christian: 1) What's the purpose of non-abridged syntax if we have an abridged one?

<StellaMitchell> PS is normative, I think

<LeoraMorgenstern> Christian, the semantics have been defined in terms of the presentation syntax, no?

<LeoraMorgenstern> So it makes sense to keep that.

<LeoraMorgenstern> But you still want something that's easier to read and write.

Christian: 2) There is a case where you need to translate/produce a RIF doc with a specific rule that is not in XML
... When you need to produce a rule set.

<DaveReynolds> Stella - the "mathematical english" is normative, the presentation syntax and it's EBNF is non-normative, the semantics actually refers to the mathematical English - at least that's how I understand it.

Chris: Not translation test cases now. But entailment test cases.
... Otherwise we would need whole suites of cases for each language.

Axel: Uneasy with +, -, *, / etc. only for numbers.
... Other languages have such symbols also for other datatypes.

<StellaMitchell> dave, the spec just says "the presenation syntax is normative" (but not the ebnf)

<AdrianP> such abreviations are quite common in nearly all programming languages

Chris: Issue if it's better than current PS.

Axel: Agree, but kind of half-way.

<DaveReynolds> Stella - yes, sorry.

<JeffP> still not clear why we need another syntax

Chris: Alternative would be to define a new syntax.

<LeoraMorgenstern> JeffP, it's because the current PS is really hard to read and write.

<LeoraMorgenstern> and we are afraid it will affect adoption of RIF.

<JeffP> LeoraMorgenstern, thanks - is PS going to be replaced by the new one?

Axel: Should we use the Abridged Syntax but mark it as not covering more than numerics?

<DaveReynolds> Leora - in that case it needs a formal definition, the current defn is ambiguous

<LeoraMorgenstern> Dave, I agree.

Chris: Will need to make status of Abridged Syntax more explict.

Axel: yes.

Leora: Discussed various more cases.

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

Leora: Quite comprehensive now.

<AxelPolleres> We should mention explicit that our abbreviations for comparison operators and arithmetic operators *only* apply to numeric datatypes, *as opposed to* the more generic use of these abbreviations in languages such as XPATH

<AdrianP> e.g. http://www.w3.org/2005/rules/wiki/Chaining_strategy_numeric-add-1

Leora: Chaining strategy test case brought in by Axel.

Axel: Dual versions, behaving differently in (naive) forward and backward engines.
... Cannot just materialize.

Leora: It would help if people make the "Discussion" slot more explicit.
... Distinguish various cases.
... Overarching argument.
... Otherwise readers can get lost.

<AdrianP> I would propose to group test cases to test suites which have a context

Chris: E.g., what Axel just described on the phone should be in the document.

Axel: Fine.

<StellaMitchell> and maybe we can use seeAlso field to reference related test cases from test cases

Sandro: "Description" vs. "Discussion" slots, with linking.

Adrian: Concept of test suite. One test case can refer to the other.

Gary: Axel's chaining strategy test cases are also legal PRD, which is good.

Leora: Stella and I translated a number of UCs.

<StellaMitchell> gary: cases that apply to BLD and PRD could be of dialect Core

Leora: Dave has a named-arg uniterm ex.
... Next week there will be more.

DTB

<ChrisW> http://lists.w3.org/Archives/Public/public-rif-wg/2008Aug/0074.html

Chris: Let's go over eds note list.

<AdrianP> Other open issues of Test Cases are e.g. the set of properties which we have define or the translation from the WIKI version into a concrete XML syntax

Chris: All shortcuts should ultimately be in one place.

Axel: Section 2.2 should explain rdf:text shortcut.
... in abridged PS, it would already be it.

Chris: Separate wiki page?

<csma> +1 for having the shortcuts and abridged syntax all in one place, and that is not DTB

Axel: No problem, but then remove it from all other places.

<AdrianP> yes, good idea

<AdrianP> YES

<AdrianP> yes

<ChrisW> ACTION: adrian to create page that documents abridged syntax [recorded in http://www.w3.org/2008/08/26-rif-minutes.html#action02]

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

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

<AdrianP> it will make it easier for programmers

<ChrisW> ACTION: apaschke to create page that documents abridged syntax [recorded in http://www.w3.org/2008/08/26-rif-minutes.html#action03]

<trackbot> Created ACTION-569 - Create page that documents abridged syntax [on Adrian Paschke - due 2008-09-02].

Axel: Predicate that takes rdf:text and language range, matching eg all "en" tags.

Dave: In SPARQL there is also a lang matcher.

Axel: A bit awkward.

Chris: Can you extract lang. tag?

Axel: Yes, but not typical use case.

Dave: Fine. Alg the same, just args different.
... Can lang tag be empty xs:string?

Axel: We agreed this should not cause any harm.

Chris: Does this change anything in the def of matching lang tag?

Axel: We could define it only for non-empty lang tags.

Dave: Same for func:lang(...)?

Axel: Dont see a problem.

Dave: Normally no need for testing non-emptiness.
... How does this affact RDF Compatiblity?

Axel: We already have different treatment there.
... treat plain literals same as those typed with xs:string.
... which is different from RDF.

Dave: According to D-entailment it is the same.
... But yours not.

Axel: Empty lang tags do not occur in RDF, so it doesnt matter.

Sandro: Syntax error in RDF.

Chris: OK. So in RDF?

Axel: Plain literals.
... So the rdf:text task force is looking into unifying them.

Sandro: No problem because rdf:text does not normally occur in RDF.

Dave/Sandro/Axel: More discussion about datatype system, performance issues, ...

Chris: Record as an issue and move on?
... Dave you are not sure?

Dave: No formal objection, need to think more about it. Feels awkward, perhaps better to raise an error instead.

Sandro: Regarding what impacts all of this might have: More on OWL than on RIF in Jena.

<ChrisW> curious

<sandro> AxelPolleres, how were you dialed in....?

<AxelPolleres> oops

<AxelPolleres> redialing

Dave: Consult with my group and come back if there is a concern.

Publicity

<sandro> http://www.w3.org/2005/rules/wiki/Publicity

Chris: Much more active Public Relations need be done.

<sandro> see for contrast http://www.w3.org/2007/OWL/wiki/Publicity

<sandro> old version.... http://www.w3.org/2005/rules/wg/wiki/Outreach

Chris: Use it to record who sent which msg where.
... What advertising has anyone done?

<DaveReynolds> Re rif:text and RDF compatibility - there isn't a problem. In the RDF spec there is the statement "The xml:lang="" form indicates the absence of a language identifier" which is therefore compatible with the proposed approach. Sorry to have diverted the discussion.

<AxelPolleres> I contacted some LP related lists... will add on the wiki

Chris: No one else? We need to start doing it NOW.
... Can we find Press Release?

<sandro> Press release: http://www.w3.org/News/2008#item132

Adrian: Some LP lists.

Axel: Sent to Enrico, the editor of the LP Newsletter.

<AdrianP> xsb-development@lists.sourceforge.net

Axel: If we have a Press Release, we can bring it in faster, otherwise 3 months.

<AxelPolleres> I did the following:

Dear all,

I would like to draw your attention to the recent drafts published by

the World Wide Web Consortium's Rule Interchange Format (RIF) working

group, towards an interchange format for rules languages:

The current specification is a collection of 6 documents published at

the following URLs:

1. RIF Use Cases and Requirements

http://www.w3.org/TR/2008/WD-rif-ucr-20080730/

2. RIF Basic Logic Dialect

http://www.w3.org/TR/2008/WD-rif-bld-20080730/

3. RIF Framework for Logic Dialects

http://www.w3.org/TR/2008/WD-rif-fld-20080730/

4. RIF RDF and OWL Compatibility

http://www.w3.org/TR/2008/WD-rif-rdf-owl-20080730/

5. RIF Production Rule Dialect

http://www.w3.org/TR/2008/WD-rif-prd-20080730/

6. RIF Datatypes and Built-Ins 1.0

http://www.w3.org/TR/2008/WD-rif-dtb-20080730/

The Working Group seeks public feedback and comments on these drafts on

the following mailinglist at the following email-address:

public-rif-comments@w3.org

with best regards,

Axel

Sandro: You have to target your audience / your community with customized versions for each mailing list.

Christian: Do not use BCC, but FWD your mailouts to Sandro.

Chris: We have to stress that BLD is in LC.

Dave: Main semweb lists.
... Jena Dev list, too

<DaveReynolds> I did publicize inside HP as well.

Gary: Send it to Oracle list and to Jess.

<AdrianP> yes there is RIF demo at the Orlando conference

<AdrianP> yes, can send it to the RuleML-2008 board

<sandro> my posting to semantic-web@w3.org http://lists.w3.org/Archives/Public/semantic-web/2008Aug/0002.html

<sandro> I'll send to swi-prolog@iai.uni-bonn.de

<sandro> semantic-web@w3.org done.

Harold: NRC and RR Board/Conf

Sandro: SWI Prolog list

<JeffP> Hi

Jeff: I'll send it to Aberdeen's rule-related projects

Mike: semweb list we covered already

<JeffP> Yuting, do you have any other suggestions?

Mike: DARPA projects

<StellaMitchell> no, no lists

<StellaMitchell> :-)

<yuting> No idea at the moment, beside the MOST project partners

Sandro: Even if we don't get much feedback, we have to present evidence that we mailed it out widely.

<StellaMitchell> I'll try

<csma> I plan to send an announcement on the JavaRules list, the Business Rules community list, the various PRR lists, the MISMO BREW list

Adrian: XSB and JBoss.

Christian: see above. Yahoo Group.

<sandro> chris: odm, common-logic,

Chris: ODM, Common Logic.

<AxelPolleres> sparql-dev?

Christian: Mark could mention it on his blog.

Chris: Other blogs. Sandro, Linked-Data community?

<AxelPolleres> public-lod@w3.org

<ChrisW> scribe for next week: GaryHallmark

AOB

<yuting> I could also send it to the ai-and-web-in-china, a google group

<DaveReynolds> bye

<JeffP> thanks, bye

<AdrianP> bye