Admin

<ChrisW> March 31 minutes: http://lists.w3.org/Archives/Public/public-rif-wg/2009Apr/att-0000/31-Mar-2009-rif-mins.html.html

<ChrisW> PROPOSED: approve minutes of last week

csma: don't remember resolution that presentation syntax should map 1:1 to xml

Sandro: design principle rather than resolution - not crisp
... e.g. literals are 1:3 mapping

<ChrisW> RESOLVED: We will use Presentation Syntax, with minor changes, with a mapping table to the XML syntax.

Sandro: from F2F7

<sandro> so a mapping could be 3:12 or whatever -- a complex rule for going from a > b to a guard expression.

ChrisW: don't see 1:1 mentioned

csma: could say we prefer 1:1 rather than refer to resolution

Sandro: shouldn't change minutes, but clarify now

<ChrisW> RESOLVED: approve minutes of last week

<sandro> I think in the meeting we were thinking it was more-or-less a WG decision to keep things 1:1, but obviously that's not really the case.

Liaison

ChrisW: added issue 91 (bounded quantifiers) to agenda
... rdf:txt discussion on email list

Jos: issues addressed, but haven't followed all email
... ready to review

Sandro: also reviewer - thinks its done

ChrisW: 1 or 2 minor things left, but shouldn't impact reviewability
... last item on agenda

<ChrisW> ACTION: josb to review rdf:text [recorded in http://www.w3.org/2009/04/07-rif-minutes.html#action01]

<trackbot> Created ACTION-725 - Review rdf:text [on Jos de Bruijn - due 2009-04-14].

csma: OMG PRR beta 2 vote on-going - AB has approved, thinking about next steps

F2F13

Sandro: ChrisW still hasn't registered
... 11 registered - usual suspects
... hope to buy lunches and snacks

<josb> sorry, our admin will not allow us to spend money, even if we do have it :(

<sandro> trackbot, help

<trackbot> See http://www.w3.org/2005/06/tracker/irc for help

Action Review

<sandro> trackbot, status?

<josb> http://lists.w3.org/Archives/Public/public-rif-wg/2009Mar/0140.html

Dave: working on applying RIF PS syntax checker to vet OWL 2 RL ruleset

Issue-67

ChrisW: string less-than issue
... broken into 2 issues
... no specific DTB predicates

<ChrisW> PROPOSED: Drop string-<, string->, string-<=, string->= from DTB, closing ISSUE-67.

ChrisW: consensus last week

<sandro> sandro: the idea is that these are just syntactic sugar for string-compare + numeric-less-than.

<ChrisW> PROPOSED: Drop string-<, string->, string-<=, string->= from DTB, closing ISSUE-67.

<AdrianP> Zakim IPcaller is me

<ChrisW> +1

<josb> no protest

<DaveReynolds> 0

<AxelPolleres> abstain (DERI)

<sandro> 0 I'm not thrilled, but okay....

0

<josb> +1

<Michael_Kifer> +1

<Harold> +1

<AdrianP> 0

<Gary> 0

<ChrisW> RESOLVED: Drop string-<, string->, string-<=, string->= from DTB, closing ISSUE-67.

<csma> +0

<ChrisW> ACTION: axel to remove string <>= from DTB [recorded in http://www.w3.org/2009/04/07-rif-minutes.html#action02]

<trackbot> Created ACTION-726 - Remove string <>= from DTB [on Axel Polleres - due 2009-04-14].

<ChrisW> ACTION: chris to close issue-67 [recorded in http://www.w3.org/2009/04/07-rif-minutes.html#action03]

<trackbot> Created ACTION-727 - Close issue-67 [on Christopher Welty - due 2009-04-14].

Issue-93

ChrisW: datatype IRIs issue, includes IRIs in general (import, prefix, annotations, datatype identifiers)

Jos: counterintuitive that 2 IRIs denote the same datatype

Sandro: don't allow equality with datatype IRIs?

<josb> xsd:string=xsd:int

Sandro: is always false

<josb> xsd:string=a

<josb> oftype("b",a)

Sandro: is true

<AxelPolleres> +1 to that being inconsistent and xsd:int=a being fine.

MKifer: rif:iri's should be uninterpreted
... use anyURI

ChrisW: plain literals refer to themselves

Jos: can live with Michael's solution

<josb> oftype("x",xsd:string)

<josb> oftype("a","http://...string"^^anyURI)

josb: different syntax for denoting datatypes

<AxelPolleres> we don't support anyURI at this point, but it could be added to DTB of course.

Sandro: concerned about equality - nice for users, but hard to implement as translation

<sandro> sandro: i had been thinking the mapping between xsd and native types was done in the translator, not in the rule engine.

<csma> A rif:iri constant must be interpreted as a reference to one and the same object regardless of the context in which that constant occurs

csma: DTB says that rif:iri constant must be interpreted as 1 object regardless of context

Jos: can be mapped to different things in different interpretations
... context is where it occurs - in formula - still same interpretation (in BLD document)

<csma> Constants in this symbol space are intended to be used in a way similar to RDF resources

csma: Make this sentence stronger?
... Make phrases more formal?

Jos: already formal in DTB document

<ChrisW> isLiteralOfType("ab"^^xsd:string, a)

<ChrisW> |=

<ChrisW> a = "http://....string"^^xsd:anyURI

<DaveReynolds> xsd:int [ rdfs:subClassOf -> xsd:integer ]

Dave: in RDFS, want to be able to make statements about datatypes

<sandro> it seems like a subproperty to me, not a subclass. :-)

<AxelPolleres> +1 to be able to talk about DTs.

Jos: nice to be able to combine RIF rules with RDF graphs

<ChrisW> "1"^^xsd:int

<ChrisW> |=

<sandro> The test case I'm concerned about is p("a"^^<a>) and <a>=xs:string |= p("a")

<ChrisW> 1 # xsd:int

<josb> sandro: this entailment does not hold

<AxelPolleres> allows to write datatype entailment rules (in a dialect allowing exiustentials in heads) it seems, e.g. rdfD1

<AxelPolleres> http://www.w3.org/TR/rdf-mt/#DtypeRules

<DaveReynolds> "1"^^xsd:int # xsd:int

<josb> xsd:string=a |= oftype("b",a)

<josb> "a"^^<a>

Josb: ^^ not evaluated

<sandro> sandro: so it's evaluated when you're doing oftype, but not when you're doing ^^.

<sandro> jos: right.,

<josb> xsd:string=a, "a"^^<a> |= oftype("a"^^<a>, xsd:string)

<josb> replace in example xsd:string w "http....string"^^anyURI

<josb> and the entailment still holds

<DaveReynolds> Right, I was talking about SWC

<AxelPolleres> rdfD1 could be emulated with skolemization in RIF BLD as: sk(?D) rdf:type ?D :- ofType(?X, ?D) ?D rd:type rdfs:Datatype.

ChrisW: what would we change to make this valid?

<sandro> sandro: sigh, yeah, I guess it's already accepted that datatypes are treated as classes (of their value space). [[ That's so broken. it means xs:hexBinary == xs:base64Binary. Sure, they are the same "class", but they are different "properties. ]]

<DaveReynolds> [[Sandro - having the same class extension does not mean xs:hexBinary == xsd:base64Binary, just means sameClassAs, can be different individuals and different property extensions. I claim it is not broken.]]

<ChrisW> xsd:int [ rdfs:subClassOf -> xsd:integer ]

<ChrisW> "1"^^xsd

<ChrisW> "1"^^xsd:int [ rdfs:type -> xsd:integer ]

<ChrisW> #

ChrisW:

Axel: makes sense
... oftype allows writing entailment rules in RIF

<AxelPolleres> http://www.w3.org/TR/rdf-mt/#DtypeRules

<ChrisW> a # ?x :- isLiteralOfType(a,?x)

<AxelPolleres> rdfD1 could be emulated with skolemization in RIF BLD as: sk(?D) rdf:type ?D :- ofType(?X, ?D) ?D rd:type rdfs:Datatype.

<ChrisW> (w/ D-Entailment)

Axel: capture finite set of inference rules in RIF

Michael: poor practice

<Michael_Kifer> datatype(?uri)

ChrisW: lots of things that BLD can't do
... understand 2 sides now
... straw poll

<csma> yes

<sandro> yes

<ChrisW> Straw: +1: Use "anyURI" in isLiteralOfType, -1: define datatype URIs to denote themselves

<josb> -1

<AdrianP> 0

<AxelPolleres> -1 (I want RIF BLD to express at least such Horn expressible inference rules that need datatype extraction over RDF, using anyURI would prevent this.)

<ChrisW> -.333333

<DaveReynolds> -0.8

<sandro> -1

-1

<Gary> -0.5

<Michael_Kifer> +1

Michael: object, because it would require quite a few changes just to accommodate 1 builtin

ChrisW: problem still exists

Michael: rif:iri's are by definition supposed to be uninterpreted - could introduce another symbol space

Jos: can define it differently

Michael: notion of datatype isn't extensible

<AxelPolleres> The built-in is anyways already an "amputed" version of SPARQL's datatype()-built-in ... if we go with anyURI, it is making even less sense to me.

Jos: hadn't thought of extensibility issue - nasty
... get rid of isLiteralOfType based on new information?
... different ways of referring to datatype inelegant

Michael: RIF pure logic, RDF evolving syntax

ChrisW: want to move on, but not lose state

<ChrisW> ACTION: mkifer to summarize objection to iris denoting themselves in email [recorded in http://www.w3.org/2009/04/07-rif-minutes.html#action05]

<trackbot> Created ACTION-728 - Summarize objection to iris denoting themselves in email [on Michael Kifer - due 2009-04-14].

Issue-95

ChrisW: list datatype issue

<AxelPolleres> the issue is: fixin the semantics on new datatypes in new rule-sets may hamper "forward"-compatibility (not sure whether that is the right term here) w.r.t. datasets not having those additional datatypes in mind, yes?

ChrisW: defer lists due to time

Issue-91

ChrisW: bounded quantifiers

<AxelPolleres> hmmm, asking myself whether the behaviour I would want could be "hidden" better in the semantics definition of isOfDatatype alone, without affecting datatypes.

ChrisW: weak support - no objections to dropping requirement

<ChrisW> PROPOSED: CORE will not have bounded quantifiers, closing ISSUE-91.

<sandro> +1

<sandro> or maybe +0

no objections

<josb> +1

<DaveReynolds> +1

<Michael_Kifer> +1

+1

<Gary> +1

<Harold> +1

<sandro> +0 they would have been nice, but it's not practical right now.

<AdrianP> +1

<ChrisW> +1

<sandro> some future Core (Core 2.0) might have it, but this Core wont have it.... We don't have time.

csma: considered deferring without closing

<ChrisW> RESOLVED: CORE will not have bounded quantifiers, closing ISSUE-91.

<ChrisW> ACTION: chris to close issue-91 [recorded in http://www.w3.org/2009/04/07-rif-minutes.html#action06]

<trackbot> Created ACTION-729 - Close issue-91 [on Christopher Welty - due 2009-04-14].

Sandro: consider starting wish list for future WG

Issue-94

rdf:text

Axel: Boris and Axel tried to resolve all open issues
... only 3 at risk notes left in document

<ChrisW> Feature At Risk #1: Usage of rtfn:

<ChrisW> Feature At Risk #2: rtfn:compare

<ChrisW> Feature At Risk #3: rtfn:length

Axel: can be emulated using existing functions

<Zakim> sandro, you wanted to ask about namespace thing

Sandro: what are different options?

<ChrisW> rdf:text editing: I see lots of undefined characters

Axel: don't see other options at this point

Sandro: prefer removing At Risk #1

<josb> I would support removing them as well

Sandro: willing to remove compare and length functions - XPath 3.0 can add them later

ChrisW: would at least document that

<josb> I actually thought there was a consensus about removing them on the rdf-text list

Sandro: obvious for XPath 3

Axel: no telecon where we formally agreed
... can resolve here from RIF side

Sandro: OWL doesn't care about builtins

<josb> I would support this

Axel: fine to drop compare and length if decided here

<josb> or ever...

Sandro: keep at risk 2 and 3 and see if we get any feedback

<sandro> sandro: let's keep length & compare as At Risk, for now.

Axel: agreed

<sandro> PROPOSED: Keep rtfn:compare and rtfn:length as AT RISK

<sandro> +1

<AxelPolleres> +1

<ChrisW> +1

<josb> 0

<Michael_Kifer> +1

<Harold> +1

<AdrianP> +1

<DaveReynolds> 0

0

<sandro> RESOLVED: Keep rtfn:compare and rtfn:length as AT RISK

<sandro> http://www.w3.org/2009/rdf-text-functions

Sandro: need Director approval for namespace - shouldn't be a problem

<ChrisW> ACTION: axel to remove at risk comment on rtfn: namespace and check with OWL WG [recorded in http://www.w3.org/2009/04/07-rif-minutes.html#action07]

<trackbot> Created ACTION-730 - Remove at risk comment on rtfn: namespace and check with OWL WG [on Axel Polleres - due 2009-04-14].

<sandro> issue-86?

<trackbot> ISSUE-86 -- rdf:text implies change to SPARQL -- OPEN

<trackbot> http://www.w3.org/2005/rules/wg/track/issues/86

<sandro> issue-87?

<trackbot> ISSUE-87 -- rdf:text document reinterprets xs:string as a subtype of rdf:text -- OPEN

<trackbot> http://www.w3.org/2005/rules/wg/track/issues/87

csma: impact on issues 86 and 87?

<ChrisW> PROPOSED: extend for 5 mins

<sandro> +1 go 5 more minutes

<csma> http://www.w3.org/2005/rules/wg/track/issues/86

<csma> http://www.w3.org/2005/rules/wg/track/issues/87

Axel: asked SPARQL WG for review

Sandro: no change required for SPARQL

<sandro> PROPOSED: Close ISSUE-86 and ISSUE-87, addressed by the current text of http://www.w3.org/2007/OWL/wiki/InternationalizedStringSpec

<josb> +1

<AxelPolleres> +1

ChrisW: try to close 86 and 87 next week

Sandro: no telecon next week

<sandro> NO TELECON NEXT WEEK. F2F13 the next day.

<ChrisW> ACTION: chris to send message about no telecon next week [recorded in http://www.w3.org/2009/04/07-rif-minutes.html#action08]

<trackbot> Created ACTION-731 - Send message about no telecon next week [on Christopher Welty - due 2009-04-14].

csma: put on agenda for F2F

ChrisW: no other business

<ChrisW> adjourn