W3C

- DRAFT -

RIF WG TeleConf Minutes

18 Apr 2006

See also: IRC log

Attendees

Present

Allen Ginsberg,  Christian de Sainte Marie, Chris Welty, Darko Anicic, Dave Reynolds, David Hirtle, Deborah Nichols, Edward Barkmeyer, Evan Wallace, Francois Bry, Frank MacCabe, Gary Hallmark, Harold Boley,  Hassan Ait-Kaci,  Ian Horrocks, Jos De Roo, Leora Morgenstern, Mala Mehrotra, Markus Kroetzsch, Michael Kifer, Paula Patranjan, Said Tabet,  Sandro Hawke, John Hall, Mohamed Zergaoui, Peter F. Patel-Schneider, Philippe Bonnard

 

Chair

Christian de Sainte Marie

Scribe

Philippe Bonnard

Contents


 

Admin

 

csma: next meeting on next Tuesday

 

<sandro> ACTION: [DONE] Christian will investigate RIF Teleconf overlapping with DAWG/SPARQL [recorded in http://www.w3.org/2006/04/04-rif-minutes.html#action04]

 

csma: proposal to accept the last minute… Proposes to postpone the acceptation.

 

<LeoraMorgenstern> about the minutes --- some action items weren't there

 

<sandro> ACTION: Leora to set up draft proposal on what we mean by FOL [recorded in http://www.w3.org/2006/04/18-rif-minutes.html#action02]

 

<sandro> ACTION: Leora to write up CSF for FOL [recorded in http://www.w3.org/2006/04/18-rif-minutes.html#action03]

 

<sandro> (these were from last time, but not recorded.)

 

 

F2F Meetings

 

csma: We'll put those in retro-actively. About F2F4, it is the last day for proposal.

 

<sandro> ACTION: Sandro set up straw-poll on F2F4 (MITRE - 3 different dates - and Peter/ISWC) [recorded in http://www.w3.org/2006/04/18-rif-minutes.html#action04]

 

Chris: straw-poll indicates the preference between the proposals.

 

<PaulaP> one vote per organization?

 

csma: a form per organization, not per individual.

 

<sandro> Deborah_Nichols: Will we have phone call-in ability for F2F3?

 

<sandro> ACTION: [DONE] Sandro to set up registration page for F2F3 [recorded in http://www.w3.org/2006/04/11-rif-minutes.html#action03]

 

<sandro> http://www.w3.org/2002/09/wbs/38457/f2f3reg/

 

<johnhall> sorry, can't get on audio yet

 

<MoZ> csma, me too, i'm interest by telcon ability

 

Paula: I think we'll have a conference phone open the whole day, but I'm not sure.

 

<SaidTabet> same question for IRC please.

 

<sandro> ACTION: Paula to check on phone-call-in capability, and if we'll have a speaker phone, and network? [recorded in http://www.w3.org/2006/04/18-rif-minutes.html#action06]

 

Liaisons

<MoZ> liaison with Xquery is difficult because of overlapping

 

 

Use cases and requirement

 

csma: About design constraint,  what is Sandro’s naming proposal for “prolog-horn” constaint ?

 

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

 

sandro: proposes "horn prolog"

 

<FrankMcCabe> I think ordered horn logic is better

 

<MoZ> +1 ordered horn clauses

 

<IanH> -1 to calling it a logic!

 

<edbark> I like Horn Prolog == Horn intersect Prolog

 

<FrankMcCabe> I think that the subset that you are interested in is sometimes called "clean prolog"

 

<Harold> Ian, so you also don't like linear 'logic' :-)

 

<Harold> What about "Sequential Horn Clauses"? Does not mention 'logic'.

 

<sandro> "Sequential Horn Clauses with Prolog Syntax"

 

<Harold> Fine with me!

 

csma: Using prolog refers to a concrete syntax

 

Francois: The name “Prolog” would suggest a full compatibility with a programming language -- something beyond what we can achieve in 1 year.

<MoZ> everybody agree to have "Horn" in the name

 

<Francois> Call it "Horn Clauses"!!!!

 

<MoZ> MoZ prefer the extension of horn as the subsetting or Prolog

 

<Francois> Can I say one more word?

 

Francois: we need something including horn logic with a prolog like syntax. Just give RIF a simple declarative Semantics.

<Francois> I do not understand Sandro's viewpoint.

 

<Francois> I would like Sandro to write down his viewpoint.

 

<Francois> I'll write my view point down.

 

Sandro: I'm not saying RIF should have procedural semantics, I'm saying we need to show how to use RIF to exchange rules in languages which do have procedural semantics.

 

<Hassan> +1

 

<Hassan> +1 on moving on

 

Sandro: Then why pick up the very old lady Prolog is and not one of the young and sexy business rule languages?

 

<sandro> Francois, because we all more-or-less know Prolog.

 

<Zakim> sandro, you wanted to respond!

 

<Francois> Sandro, who is "we", the academic crowd or the Business Rule uses?

 

<sandro> Francois, "we" is the 22 people on this call.

 

<Francois> Do we design a RIF for the 22 people on this call? Or do we want people out there to use it and make the Semantic Web a reality?

 

csma: Should RDF triples be supported by RIF ?

 

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

 

<sandro> Dave: The condition part of the a rule could match RDF triples

 

<sandro> csma: doesn't that mean RIF should have the query power of SPARQL

 

csma: If  RDF triple accepted as data, should  SPARQL query be embeddable in RIF ?

 

dave: it's more a subset of SPARQL -- the triple-match part. It leaves open whether RDF data is translated or not,  like jena rules, cwm, euler, ....

 

csma: what any consequences of accepting this requirement?

 

csma: do we mean: any RIF-compliant application, receiving a ruleset that refers to RDF triples should be able to process them?

 

Dave: cf RDF Compatibility pages --- binary predicates map to RDF triples; or a single "triple" predicate -- providing either of these would meet this requirement.

 

<sandro> (I find myself needing to see some designs before I can really understand CSMA's questions)

 

csma/sandro: as phrased this requirement is perhaps too broad to be useful in distinguishing between designs.

 

allen: maybe there's a CSF here about RDF-compatibility.

 

Dave: I agree, in my straw-man breakdown I had "RDF Compatibility" as the 3rd CSF. This was one part of that.

 

<csma> http://lists.w3.org/Archives/Public/public-rif-wg/2006Apr/0005.html

 

JosDeRoo: also RDF simple entailment rules? are they covered by this requirement? RDFS-Closure

 

<Zakim> sandro, you wanted to ask about RDF/XML Parser

 

csma: Is there a volunteer for a use case using RDF triple ?

 

sandro: I'd like to see this tied in with a use case, a scenario where RDF data is used and matters.

 

<PaulaP> there is a use case concerning access to RDF and XML data.

 

<Francois> PaulaP is right. We already have such a use case

 

<PaulaP> http://www.w3.org/2005/rules/wg/wiki/Rule-Based_Combined_Access_to_XML_and_RDF_Data contains examples of rules.

 

sandro: And I'd like to see whether or not we need an RDF/XML parser in all RIF software

 

<MoZ> http://www.w3.org/2005/rules/wg/wiki/UCR/Interchanging_Rule_Extensions_to_OWL

 

Francois: Supporting RDF triple is a central feature. If not supported, RIF is not for the semantic web. RDF semantic is beyond prolog semantic.

 

<Francois> Thanks Christian for clarifying!

 

csma: Dave should explicit the link between use case and requirement.

 

<PaulaP> the use case is not in the UCR document

 

Dave: The charter requirements about RDF are not covered by the current use case.

 

<Allen> probably in the last one: Vocabulary Mapping for Data Integration

 

<Francois> +1 with RIF accepting RDF data with its semantics.

 

<edbark> I think I will object to Sandro's rqmt for RDF parser

 

<sandro> why, edbark?

 

<Francois> Christian: the specific requirement is to work out a semantics covering Blank Nodes. This is a tough issue. But the only tough issue.

 

<edbark> +1 to accepting RIF data with its semantics

 

<MoZ> csma there is a UCR with "Requirements on the rule interchange format include semantic compatibility with OWL-DL and RDF"

 

<edbark> Sandro, it's not about syntax, it's about assertions

 

Frank: The semantic of prolog is incompatible with RDF.

 

<Francois> What means "The semantic of prolog is incompatible of RDF"? it is defined, it is different. One can make both of them compatible.

 

<Hassan> +1 with Frank

 

<IanH> Then we are done already aren't we? Surely exchanging rules is trivial if we don't care about interoperability.

 

<DaveReynolds> +1 to IanH

 

<Francois> +1 with accepting RDF assertions with their semantics

 

Francois: What kind of rules we need? Rules with a prolog like syntax working with XML, RDF and OWL data.

 

<edbark> +1 to Francois

 

Francois:  Old rule languages do not fit our needs. What is the meaning of rules? Deduction rules, constraints rules (DB, OWL, RDF), rules realizing changing (production rule).

 

<PaulaP> there are requirements regarding the different types of rules

 

<PaulaP> on the design constraints wiki page

 

<sandro> ACTION: Francois write up what he's saying on the DesignConstraints page [recorded in http://www.w3.org/2006/04/18-rif-minutes.html#action09]

 

Francois: we need a clear declarative semantics. RIF should be more abstract than traditional programming language.  I'm 99% sure no one in the Working Group wants procedural semantics for RIF. SQL's semantics leaves room for several procedural interpretations that all are correct.

 

Frank: We're not necessarily designing a new language, we're talking about interchanging existing rule languages.

 

MichaelKifer: I propose we resolve to never mention Prolog in the telecon.

 

<sandro> ...: It's an ill-formed question. If you send me some prolog text, what should I do with it? How many answers will it give before it perhaps goes into a loop where it doesn't terminate.

 

Sandro and Michael: Look at practical Use Cases like EU-Rent. The points are right but academic.

 

<sandro> Move to adjourn!

 

<Hassan> +1

 

<JosDeRoo> Francois, the rules I meant are http://www.w3.org/TR/rdf-mt/#RDFSRules

Summary of Action Items

[NEW] ACTION: Francois write up what he's saying on the DesignConstraints page [recorded in http://www.w3.org/2006/04/18-rif-minutes.html#action09]
[NEW] ACTION: Leora to set up draft proposal on what we mean by FOL [recorded in http://www.w3.org/2006/04/18-rif-minutes.html#action02]
[NEW] ACTION: Leora to Write up CSF for FOL [recorded in http://www.w3.org/2006/04/18-rif-minutes.html#action03]
[NEW] ACTION: paula check IRC too [recorded in http://www.w3.org/2006/04/18-rif-minutes.html#action08]
[NEW] ACTION: paula checking phone possibilities and speaker phone [recorded in http://www.w3.org/2006/04/18-rif-minutes.html#action07]
[NEW] ACTION: Paula to check on phone-call-in capability, and if we'll have a speaker phone, and network? [recorded in http://www.w3.org/2006/04/18-rif-minutes.html#action06]
[NEW] ACTION: Sandro set up straw-poll on F2F4 (MITRE - 3 different dates - and Peter/ISWC) [recorded in http://www.w3.org/2006/04/18-rif-minutes.html#action04]
 
[DONE] ACTION: Christian will investigate RIF Telecon overlapping with DAWG/SPARQL [recorded in http://www.w3.org/2006/04/04-rif-minutes.html#action04]
[DONE] ACTION: Sandro to set up registration page for F2F3 [recorded in http://www.w3.org/2006/04/11-rif-minutes.html#action03]
 
[End of minutes]


Minutes formatted by David Booth's scribe.perl version 1.127 (CVS log)
$Date: 2006/04/18 16:36:57 $