W3C

- DRAFT -

RDF Data Access Working group (3 Oct 2006)

3 Oct 2006

Agenda

See also: IRC log

Attendees

Present
Kendall Clark
Lee Feigenbaum
Steve Harris
Pat Hayes
Bijan Parsia
Eric Prud'hommeaux
Simon Raboczi
Andy Seaborne
Elias Torres
Regrets
Fred Zemke
Chair
Kendall Clark
Scribe
Andy Seaborne

Contents


 

nonLiteralValueTesting (process)

See minutes of 12/Sept closes nonLiteralValueTesting

Convene

RESOLUTION: to accept the minutes of 12 Sept 2006

RESOLUTION: to meet Oct 10, scribe LeeF

Action Items

Possible missing actions: http://lists.w3.org/Archives/Public/public-rdf-dawg/2006OctDec/0014.html

<kendallclark> ACTION: bijan to review FredZ Constructive mapping semantics for [recorded in http://www.w3.org/2006/10/03-dawg-minutes.html#action01]

<kendallclark> SPARQL 18 Aug [recorded in http://www.w3.org/2006/09/05-dawg-irc]

<kendallclark> DONE

<kendallclark> ACTION: [PENDING] Bijan to review FredZ 2 Aug and relate to WG issue list [DONE] [recorded in http://www.w3.org/2006/10/03-dawg-minutes.html#action02]

<kendallclark> ACTION: [PENDING] KC to review FredZ 2 Aug for issue updates [DONE] [recorded in http://www.w3.org/2006/10/03-dawg-minutes.html#action03]

<kendallclark> ACTION: Bijan to see if the Chilean's semantics paper offers any [recorded in http://www.w3.org/2006/10/03-dawg-minutes.html#action04]

<kendallclark> advice re: filters

<kendallclark> CONTINUES

<kendallclark> ACTION: bijan to write some text on the D-entailment issue [CONTINUES] [recorded in http://www.w3.org/2006/10/03-dawg-minutes.html#action05]

<kendallclark> ACTION: BijanP to propose some editorial clarification text around DATATYPE [CONTINUES] [recorded in http://www.w3.org/2006/10/03-dawg-minutes.html#action06]

<kendallclark> ericP effect: DATATYE (RDF term) => IRI | "" => xsd:string, ""@foo => error, ""^^X => X, blank node => error, IRI

<kendallclark> DONE

<scribe> ACTION: [DONE] ericP to send mail describing how [VTV] and [BTV] illustrate basic graph matching conflicts between LC1 and LC2 semantics [recorded in http://www.w3.org/2006/10/03-dawg-minutes.html#action08]

<ericP> http://lists.w3.org/Archives/Public/public-rdf-dawg/2006JulSep/0248

discussion around this to continue
... KC asks Eric to send email to push it up the stack

Convene, approve minutes, recruit scribe, etc

<bijan> AndyS, when I was reading rq24 I noticed that it seems like there has been introduced in a lot of the definition a sort of functional syntax for the algebraic operations. Is this right? Is it systmatic?

Update on pub progress for rq24 & json-sparql

<LeeF> issue was cross-site xmlhttprequest

 

EricP: pub of rq24 depends on Bijan's review - got conditional "yes"
... partial text removal (union red text)

Bijan: not conditional as such - can publish

EricP: What about the issues list comment?

Bijan: remove misleading stuff
... union was one part of this

Bijan's email: http://lists.w3.org/Archives/Public/public-rdf-dawg/2006OctDec/0001.html

<kendallclark> I'd like for it to be struck.

Kendall: wants text on costs removed.

<kendallclark> Hmm, no. I said I want explanatory, construction text removed.

<kendallclark> for the record :)

Bijan: text is not accurate and so may be confusing - better to not mention

<kendallclark> I said across the board I don't like meta-commentary in a spec.

<bijan> Why do we need to do this *now* for this pub?

<bijan> Why not later?

http://www.w3.org/2001/sw/DataAccess/rq23/rq24.html is the doc itself

<ericP> http://www.w3.org/2001/sw/DataAccess/rq23/rq24#bnodeRef

<ericP> [[

<ericP> Costs: Tableau-based reasoners (at least, the Pellet Demo example 7) rely on the current, more expressive semantics to match implications that are not in a materializable RDF graph.

<ericP> ]]

<bijan> ericP was totally garbled for me

EricP: it tries to explain why we need non-distinguished variables

Bijan: two forms of variables allows more expressive queries
... not about non-distinguished variables in patterns

<ericP> http://www.w3.org/TR/2005/WD-rdf-sparql-query-20050721/#BasicGraphPatternMatching

<ericP> vs. http://www.w3.org/TR/2006/WD-rdf-sparql-query-20060220/#BasicGraphPatternMatching

EricP: explain why there is a change in the way bNodes got implemented

Bijan: why is this about Pellet?

EricP: key is why subgraph is not sufficient

Bijan: no materialization of a graph that allows OWL-DL reasoner to integrate

<kendallclark> Hmm, so, that's 8 minutes of discussion, and I still don't hear a *candidate* for text that everyone will agree with.

PatH: suggest drop cost/benefit sections

<kendallclark> I suggest dropping the whole 'pink box'.

<kendallclark> +50

PatH: it's harmless

Kendall: Bijan's has reviewed as per the resolution
... Bijan wants it removed; Eric wants it to go as is.

Bijan: offer to send email on the topic after publication

<bijan> +1 to kendall

<bijan> My point is that the purpose of this pink box is to make a difficult issue clear, it just fails

Kendall: SW-CG want this text.
... to get community feedback

Kendall: ErciP: what else would prevent pub this week?

EricP: nothing

term distinctness for literals

EricP: nothing

[[ PatH chairs for this item ]]

Bijan: summarises issues for literals
... equality is value space or term sensitive

PatH: could have a limited amout of value sensitivity - e.g. normal forms for lexcial forms

Bijan: Fred said that there might be impls that do not retain the input form.

<ericP> how many answers will you get with DISTINCT on the non-DISTINCT dataset?:

<ericP> | type | size |

<ericP> | shoe | "10"^^xsd:integer |

<ericP> | shoe | "10"^^xsd:float |

<ericP> | shoe | "010"^^xsd:float |

AndyS: nothing about canonical forms here - this is not about reading the graph.

EricP: arises for anything to do with returning a variable in SELECT
... not about DISTINCT per se

Path: hears agreement on surface form usage - Bijan: no
... meant two literals are equal if same lex + datatype. Nothing about input

Bijan: need to do something about the loading process

AndyS: there is text around FROM for this.

<SimonR> Do we need to specify a minimum set of node equalities that MUST be evaluated, and a maximum set of equalities that may be distinguished? If they're not the same, then the implementations will possibly vary. Otherwise, we have to choose One True Equality, right?

Bijan: if this is clear, I'm fine with term distinct

<ericP> PROPSEO: DINSTINCTness depends of the surface form of the literal

<bijan> I don't find the langauge about "Dervied" on input

Bijan, I can't remember the keyword - there is something - it was a long debate a long time ago :-)

<bijan> Heh

PROPOSAL: DISTINCTness depends of the surface form of the literal (term distinct)

<kendallclark> proper search engine optimization :)

<kendallclark> We just spent nearly 20 minutes discussing informative text, so that may be overly optimistic.

Seconded: EricP

<ericP> APPROVED

<LeeF> can we get actions to get text in place for this?

RESOLUTION: DISTINCTness depends of the surface form of the literal (term distinct)

<ericP> APPROVED: Bijan abstains

 

<kendallclark> ACTION: KendallC to close formsOfDistinct issue [recorded in http://www.w3.org/2006/10/03-dawg-minutes.html#action09]

<ericP> ACTION: Bijan to propose text regarding normalization (massaging in general) while reading graphs in [recorded in http://www.w3.org/2006/10/03-dawg-minutes.html#action10]

 

<patH> Bijan requested that informative text be provided warning users about sensitivity of results to implementation stategies on graph input. (Bijan, OK??)

<scribe> ACTION: AndyS to edit text for DISTINCT = term-distinct [recorded in http://www.w3.org/2006/10/03-dawg-minutes.html#action11]

unbound variables in FILTER

[[ Kendall returns to the chair ]]

LeeF: 2 issues :: 1/ unbound variables and 2/ scope of filters

Kendall: let's talk about scope of filters

<kendallclark> http://lists.w3.org/Archives/Public/public-rdf-dawg-comments/2006Sep/0001

<

AndyS: Current 11.2 text discusses unbound variables

<ericP> http://www.w3.org/2001/sw/DataAccess/rq23/rq24#evaluation

<

<ericP> http://www.w3.org/2001/sw/DataAccess/rq23/rq24#ebv

AndyS: making 11.2.2. match 11.2

(11.2.2 is EBV)

<bijan> I'mnot ready to talk about scope of variables or filters at the moment, sorry

<kendallclark> PROPOSED: to accept the changes outlined in http://lists.w3.org/Archives/Public/public-rdf-dawg-comments/2006Sep/0001

RESOLUTION: to accept the changes outlined in http://lists.w3.org/Archives/Public/public-rdf-dawg-comments/2006Sep/0001

<ericP> APPROVED: KendallC, patH and Bijan abstaining

scope of filters

LeeF has proposed that filter scope is the group, not the BGP alone

...believes that the scope is expected to be the group

<ericP> { ?who foaf:name ?name

<ericP> FILTER (?age > 7)

<ericP> ?who foaf:age ?age }

<LeeF> I think FredZ endorsed the group scope in his attachment to

<LeeF> http://lists.w3.org/Archives/Public/public-rdf-dawg/2006JulSep/0277.html

<ericP> { ?who foaf:name ?name

<ericP> OPTIONAL { ?who foaf:age ?age }

<ericP> OPTIONAL { ?who my:age ?age }

<ericP> FILTER (?age > 7) }

<AndyS> Better example.

<ericP> 1st was just to illustrate the responsibility of FILTERing vars before they are bound

<kendallclark> LeeF: here's fred's agreement: http://lists.w3.org/Archives/Public/public-rdf-dawg/2006JulSep/0251.html

<kendallclark> This looks like the best semantics to me.

<kendallclark> I think that in general users will want to be able to sprinkle their FILTERs

<kendallclark> throughout their patterns and they should not have to worry about subtle

<kendallclark> differences depending on where they deposit a FILTER.

<LeeF> thanks, Kendall

<ericP> { ?who foaf:name ?name FILTER regexp(?name, "^Bob")

<ericP> ?who foaf:age ?age FILTER (?age > 7 }

<bijan> But how does it interact withoptional?

<kendallclark> Hmm, I'm curious whether anyone has an opinion on the next move:

<kendallclark> 1. approve Lee's tests and vote on this next week

<kendallclark> 2. approve tests and vote now

SimonR: worried if we are saying ordering is about impl hacks
... had a graph with the mathematical facts (infinite)

EricP: are we in agreement with FILTER/regex example above (filter with ?name)?

PatH: usually restrict to actual graph terms

EricP: FredZ proposed text

<bijan> http://lists.w3.org/Archives/Public/public-rdf-dawg-comments/2006Jun/0008.html

<bijan> May or may not be related

<AndyS> What exactly is the text?

AndyS:Bijan, no I don'think it is the same - they were talking about unbound, not placement

open world and other value tests

<kendallclark> http://lists.w3.org/Archives/Public/public-rdf-dawg/2006JulSep/0169.html

Tests for datatype http://lists.w3.org/Archives/Public/public-rdf-dawg/2006JulSep/0169

<bijan> mentioned in the FILTER condition to those in the pattern that is being

<bijan> filtered."""

<bijan> Sigh

<bijan> Stupid client

<kendallclark> ACTION: KendallC to put scope of filters at the top of next week's agenda [recorded in http://www.w3.org/2006/10/03-dawg-minutes.html#action12]

<patH> bijan, sounds like Jorge is talking about possible syntactic error, right? If so I agree with him.

Tests are some fixes and some for datatype()

<ericP> here's what's currently in 11

<ericP> http://www.w3.org/2001/sw/DataAccess/rq23/rq24#operandDataTypes

<ericP> simple literal denotes a plain literal with no language tag.

<ericP> http://www.w3.org/2001/sw/DataAccess/rq23/rq24#func-datatype

<ericP> Returns the datatype IRI of ltrl if ltrl is a typed literal; returns xsd:string if ltrl is a simple literal; produces an error otherwise.

<kendallclark> http://lists.w3.org/Archives/Public/public-rdf-dawg/2006JulSep/0169.html

<bijan> Just for my curiosity, have we fixed the issues raised in: http://lists.w3.org/Archives/Public/public-rdf-dawg-comments/2006Jun/0008.html ?

<kendallclark> ACTION: PatH to review the proposed tests in http://lists.w3.org/Archives/Public/public-rdf-dawg/2006JulSep/0169.html and say yay or nay [recorded in http://www.w3.org/2006/10/03-dawg-minutes.html#action13]

<bijan> E.g., """for some graph patterns P it is not always the case that {P . P} gives the same result as { P }."""

Open World tests 2006JulSep/0180

<kendallclark> ACTION: EricP to review the tests in http://lists.w3.org/Archives/Public/public-rdf-dawg/2006JulSep/0180.html and say yay or nay [recorded in http://www.w3.org/2006/10/03-dawg-minutes.html#action14]

<scribe> ACTION: Bijan review rq24 against http://lists.w3.org/Archives/Public/public-rdf-dawg-comments/2006Jun/0008.html [recorded in http://www.w3.org/2006/10/03-dawg-minutes.html#action15]

Protocol spec

Kendall: does not have time to chair and edit the protocol spec
... asked Elias and Lee to become editors
... and they have agreed

<bijan> kendall, perhaps send an annouce of the appt to the mailing list?

<ericP> sounds good to me

Summary of Action Items

[NEW] ACTION: AndyS to edit text for DISTINCT = term-distinct [recorded in http://www.w3.org/2006/10/03-dawg-minutes.html#action11]
[NEW] ACTION: Bijan review rq24 against http://lists.w3.org/Archives/Public/public-rdf-dawg-comments/2006Jun/0008.html [recorded in http://www.w3.org/2006/10/03-dawg-minutes.html#action15]
[NEW] ACTION: Bijan to propose text regarding normalization (massaging in general) while reading graphs in [recorded in http://www.w3.org/2006/10/03-dawg-minutes.html#action10]
[NEW] ACTION: Bijan to see if the Chilean's semantics paper offers any [recorded in http://www.w3.org/2006/10/03-dawg-minutes.html#action04]
[NEW] ACTION: EricP to review the tests in http://lists.w3.org/Archives/Public/public-rdf-dawg/2006JulSep/0180.html and say yay or nay [recorded in http://www.w3.org/2006/10/03-dawg-minutes.html#action14]
[NEW] ACTION: KendallC to close formsOfDistinct issue [recorded in http://www.w3.org/2006/10/03-dawg-minutes.html#action09]
[NEW] ACTION: KendallC to put scope of filters at the top of next week's agenda [recorded in http://www.w3.org/2006/10/03-dawg-minutes.html#action12]
[NEW] ACTION: PatH to review the proposed tests in http://lists.w3.org/Archives/Public/public-rdf-dawg/2006JulSep/0169.html and say yay or nay [recorded in http://www.w3.org/2006/10/03-dawg-minutes.html#action13]
 
[PENDING] ACTION: bijan to write some text on the D-entailment issue [recorded in http://www.w3.org/2006/10/03-dawg-minutes.html#action05]
[PENDING] ACTION: BijanP to propose some editorial clarification text around DATATYPE [recorded in http://www.w3.org/2006/10/03-dawg-minutes.html#action06]
 
[DONE] ACTION: Bijan to review FredZ 2 Aug and relate to WG issue list [recorded in http://www.w3.org/2006/10/03-dawg-minutes.html#action02]
[DONE] ACTION: EliasT to follow up w/ Andy on "the idiom" for plain literals/string literals [recorded in http://www.w3.org/2006/10/03-dawg-minutes.html#action07]
[DONE] ACTION: ericP to send mail describing how [VTV] and [BTV] illustrate basic graph matching conflicts between LC1 and LC2 semantics [recorded in http://www.w3.org/2006/10/03-dawg-minutes.html#action08]
[DONE] ACTION: KC to review FredZ 2 Aug for issue updates [recorded in http://www.w3.org/2006/10/03-dawg-minutes.html#action03]
 
[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.127 (CVS log)