RIF Telecon 24 March 2009 -- 24 Mar 2009

<ChrisW> Harold, I've put the new schema on the agenda

<Harold> Chris, thanks.

Admin

<ChrisW> PROPOSED: approve last week's minutes

<ChrisW> http://lists.w3.org/Archives/Public/public-rif-wg/2009Mar/att-0035/03-rif-minutes.html

<ChrisW> RESOLVED: approve last week's minutes

Liason

OWL WG has taken on board a RIF request on ???

Axel: SPARQL discussions on proposed features in progress

<AxelPolleres> http://www.w3.org/2009/sparql/wiki/Category:Features

F2F13

Chris: Recommend F2F attendees get hotel reservations in ASAP

Action Review

All actions continued

Liaison

New XML Schema

Harold: Finished schema but changkai not present - return to later in meeting

Issue-92

<ChrisW> PROPOSED:* To keep the separation between preds, funcs, and individuals, but pred, func, external symbols can have multiple arities. Closing ISSUE-92 [5].

<ChrisW> PROPOSED: To keep the separation between preds, funcs, and individuals, but pred, func, external symbols can have multiple arities. Closing ISSUE-92 [5].

<sandro> +1

<AxelPolleres> +1

<DaveReynolds> +1

<josb> 0

<ChrisW> +1

<mdean> +1

<PaulV> 0

<gary> 0

<Harold> 0

<Michael_Kifer> +1

<csma> 0

<ChrisW> RESOLVED: To keep the separation between preds, funcs, and individuals, but pred, func, external symbols can have multiple arities. Closing ISSUE-92 [5].

<ChrisW> ACTION: chris to close issue 92 [recorded in http://www.w3.org/2009/03/24-rif-minutes.html#action01]

<trackbot> Created ACTION-716 - Close issue 92 [on Christopher Welty - due 2009-03-31].

Issue-80

<DaveReynolds> http://lists.w3.org/Archives/Public/public-rif-wg/2009Mar/0076.html

CSMA: options are from Dave's email - see agenda

- From the minutes of last week [1], the only proposed resolution with only positive opinions according to the polls is the following (but Dave and Jos were not present and did not state an opinion about that option in there emails) * Option (1) would be opposed (-1) by Adrian, and option (0) by sandro; * option (2), preferred by Dave, has mildly negation opinions (0.5) from Gary and Jos, and slightly negative opinion (0.1) from Michael; * option (3), pr

Proposal was: Have xpath-style-equals, xpath-style-not-equals, AND pred:literal-not-equal for non-identical-literals (eg for OWL RL

<josb> I would object to 2

<josb> see also the Mar 3rd minutes

<ChrisW> PROPOSED:* Have xpath-style-equals, xpath-style-not-equals, AND pred:literal-not-equal for non-identical-literals (eg for OWL RL). Closing ISSUE-80 [6].

<ChrisW> PROPOSED: Have xpath-style-equals, xpath-style-not-equals, AND pred:literal-not-equal for non-identical-literals (eg for OWL RL). Closing ISSUE-80 [6].

<AxelPolleres> you don't mean new predicates do you?

Christian: this is a combination of options 1 and 3

<josb> unless someone makes a compelling case, I will object

Chris: to summarize.... no I cant

Dave: addition is xpath-style-no-equals

Christian: proposed only from the polls last week as the least opposed solution

Axel: 2 new predicates - would they be extensible or not? Christian: they would keep extensibility

<josb> NO

<josb> I was not there last week

<AxelPolleres> wasn't there either

<gary> 1 is xpath-equal to 1.0 but not literal-equal

<csma> option "(2)" pred:literal-equal "as is" --- literal-equal(x,y) iff literal(x) and literal(y) and x=y

<josb> 2009-01-01T01:00 GMT neq 2009-01-01T02:00 GMT+1

<josb> 2009-01-01T01:00 GMT not-identical 2009-01-01T02:00 GMT+1

<josb> 2009-01-01T01:00 GMT datetime-equal 2009-01-01T02:00 GMT+1

<AxelPolleres> but they are '=', aren't they?

<sandro> I think have literal-not-equals is important, not literal-not-identical.

<AxelPolleres> literal-equal is currently "=" plus a value space check

<AxelPolleres> so: literal-equal ( 2009-01-01T01:00 GMT 2009-01-01T02:00 GMT+1 ) holds

<AxelPolleres> I don't know what not-identical is.

<ChrisW> identical is rif:=

<ChrisW> equality is datatype-equality e.g. with numerics and dates

<AxelPolleres> http://www.w3.org/2005/rules/wiki/DTB#pred:literal-equal

<sandro> Sandro: I *think* for owl RL we need non-identity in the value space of literals, not numeric-not-equal.

<sandro> chris: Dave's ruleset needs literal-non-identity

<sandro> literal-identical, literal-non-identical, literal-equal, literal-not-equal

<AxelPolleres> literal-identical is redundant

<AxelPolleres> literal-equal had promotion?

<josb> yes

Literal-identical is literal-identical or literal-qual with existing RIF?

<AxelPolleres> literal-not-equal means unequal incl. promotion

<AxelPolleres> literal-not-identical is what is now literal-not-equal in the spec

<sandro> gary wants literal-equals

<ChrisW> literal-not-identical: use case for OWL-RL ruleset

<ChrisW> literal-identical: symmetric with above

<sandro> axel: What's in the current draft as literal-equal and literal-non-equal is what we're currently talking about as literal-identical and literal-not-identical.

<AxelPolleres> BUT: literal-not-equal (incl. promotion) would back up XPAths !=

<sandro> chris: need to verify that OWL is using identity, not xpath-style-equality.

<DaveReynolds> Jos - are you sure it is doing so for xsd:dateTimeStamp?

Jos: OWL is using identity as we use in RIF

<sandro> PROPOSED: include literal-not-identical in DTB. true for two literals that map to the same point in the value space. false if different data type. false if either is not a literal. false if same type and different point in value space

<josb> yes, it always uses identity

<AxelPolleres> literal-not-identical would be jsut obtained by renaming the current http://www.w3.org/2005/rules/wiki/DTB#pred:literal-not-equal

<sandro> PROPOSED: include literal-not-identical in DTB. false for two literals that map to the same point in the value space. true if different data type. false if either is not a literal. true if same type and different point in value space

<AxelPolleres> the proposal amounts to "PROPOSED: rename pred:literal-not-equal to pred:literal-not-identical", correct?

<ChrisW> ACTION: jos to check impact of pred:literal-not-identical to OWL-RL ruleset [recorded in http://www.w3.org/2009/03/24-rif-minutes.html#action02]

<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: josb to check impact of pred:literal-not-identical to OWL-RL ruleset [recorded in http://www.w3.org/2009/03/24-rif-minutes.html#action03]

<trackbot> Created ACTION-717 - Check impact of pred:literal-not-identical to OWL-RL ruleset [on Jos de Bruijn - due 2009-03-31].

<AxelPolleres> I don't see anything changed from the current status, except a renaming proposal from -equal to -identical

rdf:text

<josb> URI?

- The OWL WG would like to move rdf:text to LC [10]. What shall we do about ISSUE-86 [8] (rdf:text implies changes to SPARQL) and ISSUE-87 [9] (rdf:text re-interprets xs:string as a subtype of rdf:text)?

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

<AxelPolleres> I didn't have a handle on it recently, sorry.

<AxelPolleres> I can take an action to get up-to-date on the issues and status by one but nexty teleconf. (will be on a review meeting next week)

Sandro: lots of editorial notes still in this spec
... call for reviewers

Chris: needs to be discussed prior to closing - reviewer will be Jos

<ChrisW> ACTION: sandro to review rdf:text for RIF [recorded in http://www.w3.org/2009/03/24-rif-minutes.html#action04]

<trackbot> Created ACTION-718 - Review rdf:text for RIF [on Sandro Hawke - due 2009-03-31].

<Harold> http://www.w3.org/2005/rules/wiki/Core#Appendix:_XML_Schema_for_RIF-Core

<Harold> http://www.jdrew.org/rif/core/wd2/absquared.rif

<scribe> Chair: are we on Agendum 5?

<Harold> http://www.jdrew.org/rif/core/wd2/ex7CounterExample.rif

Lists

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

[13] http://www.w3.org/mid/20090317032509.51624560%2540kiferserv

<josb> seems fine

Sandro: thought we wanted a list for Core e.g. with standard builtins, works with Java for PRD, and could be implemented in BLD etc with Prolog lists if desired

<Harold> Sandro, we did have discussions about 'built-in' lists, but then converged to go for general lists that can be non-ground.

<josb> I suppose we would not have lists in the head in Core?

<josb> How could we then define append?

Christian: not sure that proposal is what PRD required... eg X is a list and is X empty and membership tests and counts

Michael: these can all be done

<Harold> Jos, append would not need to be user-definable in Core (although in BLD): we could have a concatenation built-in on top of Michael's revised list proposal.

Christian: variables in lists cannot be done in PRD

<Gary> variables should be ok as long as safety is respected

<josb> OK, that's what I thought

Christian: why in PRD would a list need to be a different type versus another datatype

<sandro> csma was imagining lists being external, not giving in a rif ruleset.

<sandro> csma: If we have no variables inside lists inside core, maybe that's equivalent.

<sandro> if Seq(1 ?x ?y 3) == ?z then ....

<sandro> sandro: this is equivalent to putting Function Terms in core.

<sandro> kifer: true, but not if we have safety.

<josb> but you need to disallow lists with vars in head

Isn't this more of a regexp type of "list"?

Core-ifying SWC

<sandro> gary: note that if you have trees, you'd need to recurse down to match up the variables, like unifying terms.

Jos: new potential issue...

<josb> http://lists.w3.org/Archives/Public/public-rif-wg/2009Mar/0021.html

Jos prefers disallow using rdfs:Resource in the rules and in RDF triples that are not of the form xxx rdf:type rdfs:Resource in the embedding

<AxelPolleres> sigh.

Chris: AOB?
... Adjourn

- DRAFT -

RIF Telecon 24 March 2009

24 Mar 2009

Attendees

Contents