Admin

<ChrisW> Minutes from last meeting: http://lists.w3.org/Archives/Public/public-rif-wg/2010Jun/att-0006/08-rif-minutes.html

<ChrisW> PROPOSED: accept last meeting minutes

<ChrisW> RESOLVED: accept last meeting minutes

Liaison

Sandro: SPARQL 1.1 adds entailment regimes
... specifies what it means to use SPARQL with OWL 2, RDFS, ..., RIF
... FPWG a month or so ago
... should be reviewed by at least one other RIF WG member
... need official RIF WG review
... only a page or 2, but fairly mathematical

<ChrisW> ACTION: Mike_Dean to review sparql entailment regimes on sparql [recorded in http://www.w3.org/2010/07/13-rif-minutes.html#action01]

<DaveReynolds> http://www.w3.org/TR/sparql11-entailment/#id35811294

<ChrisW> ACTION: MDean to review sparql entailment regimes on sparql [recorded in http://www.w3.org/2010/07/13-rif-minutes.html#action02]

<trackbot> Created ACTION-1033 - Review sparql entailment regimes on sparql [on Mike Dean - due 2010-07-20].

ChrisW: RDFRDB?

Sandro: Eric Prud'hommeaux sent email to various folks a couple weeks ago
... define a standard mapping from RDBMS to RDF graph with minimal configuration
... use RIF for other mappings - frames to frames
... provided RIF examples recently
... propose as RIF test cases - e.g. changing URLs
... also other proposals
... remains to be seen how well other RDF2RDB WG members will accept this - lots of folks would like to do everything in SQL

Sandro: want these transformations for other cases as well

Sandro: eGov is thinking of starting at WG on government linked data
... still working on proposed charter
... see eGov IG email list

Actions

<ChrisW> close action-1030

<trackbot> ACTION-1030 Setup poll for RIF WG meeting for Jun 17th or 18th, to talk about XML-Data and RIF-in-RDF drafts for publication on 22nd. closed

close action-1029

<trackbot> ACTION-1029 Contact Bijan re: AC support closed

close action-1028

<trackbot> ACTION-1028 Contact Jeff Pan re: AC support closed

close action-1026

<trackbot> ACTION-1026 Talk to Elisa re: AC support closed

close action-1025

<trackbot> ACTION-1025 Talk to Mark Proctor re: AC support closed

close action-1024

<trackbot> ACTION-1024 Talk to OMG re: AC support closed

close action-1023

<trackbot> ACTION-1023 Talk to Ed Barkmeyer re: AC support closed

close action-1022

<trackbot> ACTION-1022 Talk to Guido re: AC support closed

close action-1021

<trackbot> ACTION-1021 Talk to Ken for Mitre re: AC support closed

close action-1019

<trackbot> ACTION-1019 Contact Mohamed Zergaoui re: AC suport closed

close action-1018

<trackbot> ACTION-1018 Contact Michael Sintek re AC support closed

close action-1017

<trackbot> ACTION-1017 Email Luis Polo re AC support closed

close action-1016

<trackbot> ACTION-1016 Email Enrico Franconi re AC support closed

close action-1015

<trackbot> ACTION-1015 Email Markus Krotzsch re AC support closed

close action-1014

<trackbot> ACTION-1014 Contact Clevand Clinic to encourage AC support closed

ChrisW: do we have a record of how AC voted?

Sandro: no - about 15 support - don't remember exact number

close action-1013

<trackbot> ACTION-1013 Fix the reference to Proposed Rec in the Media Type Registraiton, in Core closed

<ChrisW> christian?

<csma> yes

<csma> just found the number

close action-968

<trackbot> ACTION-968 Document an rdf syntax for rif closed

ChrisW: open actions will be closed when WG ends

csma: sent agenda from private address, so it didn't go to public email list

XML Data

csma: posted remaining issues
... needs to be finished
... RDF plain literals - get rid of name match
... Jos suggestions
... choice of syntax - @ vs attribute - not sure if allowed in IRIs
... dealing with case where elements/attributes not in namespace - cannot use rif:iri's - currently use strings, but not sure if this is allowed
... document must be completed with more and better examples, reviewed in more depth, etc.
... don't expect major changes in content
... Jos and Mike didn't find any showstoppers
... plan to work on this next week - proposed solutions for above issues - discuss at next meeting
... don't currently need help - will need later for review, implementation, etc.
... expect new version in 2 weeks
... would like Michael to look at it also

MichaelKifer: looked at version before Jos's comments, but didn't formally review

csma: incorporated Jos' and Mike's comments - haven't yet gotten feedback from Jos

MichaelKifer: will take a look

Sandro: use of special "no namespace" namespace rather than strings?

csma: most obvious solution would be to change rif:iri to allow local names
... but potentially opens up a can of worms
... don't like "no namespace" namespace
... only preferences - need to find most practical and least intrusive solution
... do other specs have this problem?

Sandro: RDF and OWL don't deal with XML - others model in full complexity (might or might not have namespace)

csma: would prefer using ony rif:iri to address children and attributes, instead of having to support both rif:ri and xs:strings just because of the nonamespace case

RIF in RDF

Sandro: 2 open issues and omission from FPWG
... both about what modeling is done: use type arcs, what to name various properties
... mapping between RIF syntactic structures and RDF graphs
... should these include rdf:type arcs, e.g. for RIF formula
... or could just list properties
... seems to be purely stylistic
... left out rdf:type arcs for now - recorded as open issue

ChrisW: define semantics so that this is implied

Sandro: could provide OWL ontology

mdean: or RDFS domain/range

Sandro: have to name some things differently than XML syntax, e.g. distinguish AND and OR
... made up names
... 3rd thing: importing or asserting RIF from RDF
... use case: Facebook RDF data using vocabulary unlike most others, e.g. using strings rather than URIs
... can be mapped using RIF rules - but how does one find these?
... triples in RDF should specify RIF rules to use
... SPARQL also needs for entailment regime
... would add to SWC if not too late

ChrisW: what happens when you dereference URIs?

<DaveReynolds> For the record, as Sandro says, I prefer use of type arcs to enable consistent property naming and for consistency with other specs, e.g. OWL. Easy to say which are require and which optional.

Sandro: have so far only handled RIF import RDF case, not RDF import RIF

<DaveReynolds> +q

Sandro: how will RDF find rules
... extra parameter beyond owl:imports to specify profile
... possibly property that implies profile

ChrisW: not simple, seems very tricky

DaveReynolds: shouldn't do the same job
... need RDF imports

Sandro: not discussed at RDF NS workshop
... little awareness of RIF
... not in final list
... SPARQL has in entailment WDs
... part of graph

ChrisW: seems like it should be part of RDF

Sandro: semantics and namespace more important than which document it's in

Test

ChrisW: Stella not here
... no active development

Christian: no need to publish until updated

Sandro: can't run Google Analytics on W3C site, to tell who's reading documents and how they got there

<ChrisW> ACTION: Chris to contact Stella about test cases status [recorded in http://www.w3.org/2010/07/13-rif-minutes.html#action03]

<trackbot> Created ACTION-1034 - Contact Stella about test cases status [on Christopher Welty - due 2010-07-20].

Primer

Leora: late, but have draft on Wiki - bulk of it is there

<lmorgens> http://www.w3.org/2005/rules/wiki/Primer

Leora: had hoped to have Harold and Chris review

ChrisW: meeting to follow up on Friday 11am Eastern

Leora: not sure if reached right level of detail and approachability
... e.g. section on reasoning - brief discussion of semantic issues, approximation
... good to get feedback on sections like this

ChrisW: concrete example from Facebook use case

Sandro: would like to make concrete
... properties associated with page, rather than object - and values are strings rather than URIs

<Harold> http://developers.facebook.com/docs/opengraph

Sandro: can see whether appropriate for Primer

<DaveReynolds> Sounds like a good use case for the primer given interest in OGF and emphasises RDF connection.

Leora: current examples stick to one theme, but another might be helpful given that RIF users will also have to deal with issues like this

ChrisW: will review this week for Friday's meeting

Sandro: what syntax does the primer use?

Leora: antecedent preceding consequent
... Chris wants to change slot value
... or use both, with hide/show button - simplified vs. presentation

Sandro: prefers button

<sandro> Here's one hide/show bit: http://www.w3.org/2009/02/multisyntax-support/example-1.html

ChrisW: should provide XML for each example

Sandro: totally in favor of having every widely used syntax

<csma> +1 to having multiple syntaxes in the primer

Sandro: very helpful in OWL primer
... but a lot of work

ChrisW: can be done at end
... once content is stable

Harold: would prefer Simplified RIF Syntax to RIF Simplified Syntax (another RSS)

Leora: Will do

<sandro> -1 to "Simplified RIF Syntax"

Leora: haven't checked parens - will do later today

Harold: eliminate extra spaces

Sandro: Simplified RIF Syntax implies something official

ChrisW: perhaps RIF Primer Syntax

<sandro> "RIF Primer Syntax" is fine.

<csma> why do you need a name?

Sandro: rest of world calls this tutorial not primer
... other W3C specs also use primer

csma: primer not necessarily tutorial
... just examples, not necessarily complete

ChrisW: send naming suggestions by email
... Any other Business?

<sandro> (I agree, CSMA -- I guess I don't know which this is.)

ChrisW: Meeting in 2 weeks, July 27?

<DaveReynolds> Date OK for me

ChrisW: meeting in 2 weeks, same agenda as today
... adjourned

Adrian joined late, in UK