W3C

- DRAFT -

SPARQL Working Group Teleconference

29 Mar 2011

Agenda

See also: IRC log

Attendees

Present
AxelPolleres, cbuilara, bglimm, LeeF, kasei, OlivierCorby, MattPerry, SteveH__, pgearon, alepas, AndyS, Souri_, NicoM
Regrets
Chime, NickH, sandro
Chair
LeeF
Scribe
bglimm

Contents


<trackbot> Date: 29 March 2011

<LeeF> scribenick: bglimm

Admin

<LeeF> PROPOSED: Approve minutes at http://www.w3.org/2009/sparql/meeting/2011-03-22

<LeeF> RESOLVED: Approve minutes at http://www.w3.org/2009/sparql/meeting/2011-03-22

<LeeF> Next regular meeting: 2011-04-05 @ 15:00 UK / 10:00 EST (scribe: Axel or Alex)

<AxelPolleres> comments page should be up-to-date

LeeF: We have some comments, mostly under control

<LeeF> http://www.w3.org/2009/sparql/wiki/Comments

<AxelPolleres> some are unasigned, still

LeeF: maybe spend some tome on not in, in
... SPARQL implementations currently have no understanding of datatypes, which can result in unintuitive results for comparissons
... should SPARQL prescribe some understanding for core datatypes in comparrisson operators
... Andy, Steve, should we look into that?

SteveH: Seems like an improvement, but not full understanding

AndyS: The comment is related to 1.0 stuff and not specific to 1.1

<AndyS> It will change a basic, unextended SPARQL 1.0 query processor. (we should have done it last time but that makes it a change)

LeeF: Not much enthusiasm for this topic, so lets not spend too much time on it

AxelPolleres: We are mostly up-to-date

<SteveH> would be it be sufficient to recommend that SPARQL 1.1 processors should handle all the datatypes so that...

<LeeF> http://lists.w3.org/Archives/Public/public-rdf-dawg-comments/2011Mar/0018.html

<AndyS> While sensible, it's technically a change. Not sure if its in the 1.0 test suite or not.

<SteveH> SHOULD or something

AxelPolleres: regarding the comments

Last Call Status

<LeeF> http://www.w3.org/2009/sparql/wiki/To_Last_Call

LeeF: Lets go through the documents and editors correct me if I am wrong with something
... Query has still some editorial comments and aggregates algebra section
... has still some things that can be improved

SteveH: Not much knew from me

AndyS: I did some changes for RDF merge and wait for Axel's second part of the review

LeeF: If I had some time, shoud I rather work on the protocoll or review query?

AndyS: We had already three reviews, so I think protocol is more important to get done

SteveH: Same from my side

<kasei> was I meant to start that review yet? I thought I was waiting on somebody to ping me on that?

LeeF: Update had some work done regarding Axel's review, Andy's review is still to be addressed

<LeeF> kasei, ah, i did not realize that

AxelPolleres: There is nothing that cannot be resolved by us. Andy suggested some restructuring to make the distinction between the formal and informal part
... that has not been done yet
... I'll sync up with Paul and Alex for that. We are not too far from LC

LeeF: I think Greg can go ahead with his review

Axel: I worked in parallel to Axel, so now we have to resolve CVS conflicts and get an overview again
... seems Axel has done a lot of stuff, so most might be addressed

<kasei> ok

AxelPolleres: Points that are open are marked in my email answer to Andy

<LeeF> kasei, thanks

<kasei> am out of the country now, but will try to review it soon

AxelPolleres: some things are left open because I wait for conirmation from Paul

LeeF: Protocol, nothing new, same for Service Descriptions

<kasei> correct

LeeF: RDF Dataset/HTTP Protocol, we have to look at the name of the doc shortly
... Kjetil's comments still have to be considered

bglimm: d-entailment updated for ent. regimes and section added for property paths

<LeeF> ACTION: Matt to look at new d-entailment text [recorded in http://www.w3.org/2011/03/29-sparql-minutes.html#action01]

<trackbot> Sorry, couldn't find user - Matt

LeeF: Matt, can you look at the D-Entailment section?

<LeeF> ACTION: Matthew to look at new d-entailment text [recorded in http://www.w3.org/2011/03/29-sparql-minutes.html#action02]

<trackbot> Created ACTION-422 - Look at new d-entailment text [on Matthew Perry - due 2011-04-05].

Matt: Yes

LeeF: Federated Query is waiting on Axel and myself to finish the review
... other documents nothing new
... Anything important for LC?

Name of the RDF dataset protocol specification

(silence)

LeeF: We realised that the name might not be appropriate

<LeeF> PROPOSED: Rename the datset protocol to the SPARQL 1.1 Graph Store HTTP Protocol

LeeF: It is about managing graph stores. So we could change dataset to graph store
... Chime is ok with that
... any objections to that change?

<AndyS> seconded

<NicoM> +1

<AxelPolleres> +1

<Souri> +1

<LeeF> RESOLVED: Rename the datset protocol to the SPARQL 1.1 Graph Store HTTP Protocol

<AxelPolleres> Note that we need to propagate this change to other docs referring to that one!

tests

<LeeF> close ACTION-421

<trackbot> ACTION-421 Look through test cases and provide a summary by next TC closed

LeeF: Olivier looked through the test cases and where we stand

<kasei> I reference it in the SD doc

<kasei> will change

me too

<LeeF> AxelPolleres: we need to make sure that all other documents update the name of the http protocol document

<LeeF> http://lists.w3.org/Archives/Public/public-rdf-dawg/2011JanMar/0501.html

LeeF: Hope we can approve some tests
... Olivier ran the tests with his implementation

<LeeF> Corese

LeeF: two areas where we miss test cases

<AndyS> See http://www.w3.org/2009/sparql/docs/tests/data-sparql11/manifest-all.ttl ARQ gets: Tests = 332 : Successes = 298 : Errors = 9 : Failures = 25

LeeF: for the IF function and scoping? for zero length paths
... AndyS, can we get that covered

AndyS: I was hoping for WG support for this

LeeF: Matt, would you mind to come up with a test that covers the missing property path features?

<LeeF> ACTION: Matthew to include a test on nodes in path of length zero come from specified named graph (e.g. graph <g1> {?x <p>* ?y}) [recorded in http://www.w3.org/2011/03/29-sparql-minutes.html#action03]

<trackbot> Created ACTION-423 - Include a test on nodes in path of length zero come from specified named graph (e.g. graph <g1> {?x <p>* ?y}) [on Matthew Perry - due 2011-04-05].

Matt: I can do that

<LeeF> ACTION: Lee to follow-up and make sure we get IF() tests [recorded in http://www.w3.org/2011/03/29-sparql-minutes.html#action04]

<trackbot> Created ACTION-424 - Follow-up and make sure we get IF() tests [on Lee Feigenbaum - due 2011-04-05].

LeeF: We still have some empty directories
... I fixed the manifest now.
... Olivier picked up on negative syntax tests, should we keep these types?
... Do we have similar types for the positive tests?
... Any opinions?

(silence)

<kasei> is it in a different namespace?

<kasei> the new 1.1 namespace?

<LeeF> mf:NegativeSyntaxTest

<LeeF> mf:NegativeSyntaxTest11

<LeeF> mf:NegativeUpdateSyntaxTest11

<kasei> mf is the old dawg namespace, then?

<LeeF> @prefix mf: <http://www.w3.org/2001/sw/DataAccess/tests/test-manifest#> .

<kasei> if it's the same namespace, I'd prefer keeping the '11'

AndyS: I suggest the negative syntax tests without 1.1 should be changed

LeeF: Should we move all syntax tests into one directory?

<AxelPolleres> I didn't add those classes yet, I am afraid (the ones with 11)

AndyS: The syntax tests should have coverage even if hey are scattared in different directories

AxelPolleres: Do we need NegativeSyntaxTest11 and NegativeUpdateSyntaxTest11?

<kasei> think it needs to either be just negativesyntaxtest11 or be explicit by changing NegativeSyntaxTest to NegativeQuerySyntaxTest

<kasei> prefer the latter

AxelPolleres: Is that not clear from the fact that the test is an update test or a query test?

AndyS: I find it clearer the way it is.

AxelPolleres: Do we need the same for PositiveSyntax...Test?

AndyS: I think we have that already.
... Yes, and that's the only way to distinguish them, so we have to keep it

<AxelPolleres> I need an action to add those new types to mf: and to README.html

AndyS: For the syntax tests the type is important to distinguish them

LeeF: The ones in Aggregates without the 11 have to be updated. I'll do that now

<AxelPolleres> ACTION: Axel to add those new types to mf: and to README.html [recorded in http://www.w3.org/2011/03/29-sparql-minutes.html#action05]

<trackbot> Created ACTION-425 - Add those new types to mf: and to README.html [on Axel Polleres - due 2011-04-05].

<AxelPolleres> ... new types: Positive/NegativeSyntaxTest11 and Positive/NegativeUpdateSyntaxTest11 yes?

LeeF: Do you know whether that is the only place?

AndyS: Yes.

LeeF: Ok, then I updated that

<LeeF> ACTION: Andy to change the @prefix : prefix in the syntax directories to use an absolute URI [recorded in http://www.w3.org/2011/03/29-sparql-minutes.html#action06]

<trackbot> Created ACTION-426 - Change the @prefix : prefix in the syntax directories to use an absolute URI [on Andy Seaborne - due 2011-04-05].

LeeF: We are not consistent in the update tests for specifying the data

AxelPolleres: We don't really need new types for the query evaluation tests

<AxelPolleres> :QueryEvaluationTest vs :QueryEvaluationTest11 ?

AndyS: We make sure that when you execute the test suite, you do 1.1.

<LeeF> ut:graphData [ ut:graph

<LeeF> ut:graphData [ ut:data

LeeF: ut:data vs. ut:graph, which one should be used?

AxelPolleres: data is just for the default graph

LeeF: Is the second of my example incorrect?

<AxelPolleres> http://www.w3.org/2009/sparql/docs/tests/README.html#updateevaltests

<LeeF> ACTION: Lee to clean up occurrences of ut:graphData [ ut:data [recorded in http://www.w3.org/2011/03/29-sparql-minutes.html#action07]

<trackbot> Created ACTION-427 - Clean up occurrences of ut:graphData [ ut:data [on Lee Feigenbaum - due 2011-04-05].

LeeF: extra prefixes don't harm, could be cleaned up
... same for duplicate tests
... delete/insert, we know about the blank node in the template issue
... should now be a negative syntax test

<LeeF> ACTION: Lee to fix the delete-insert queries that should be negative syntax tests in delete-insert [recorded in http://www.w3.org/2011/03/29-sparql-minutes.html#action08]

<trackbot> Created ACTION-428 - Fix the delete-insert queries that should be negative syntax tests in delete-insert [on Lee Feigenbaum - due 2011-04-05].

LeeF: Olivier identified six ent. test cases that have a mistake

bglimm: I'll check that

LeeF: Axel, Olivier had some comments for the readMe document, can you address those?

<AxelPolleres> ACTION: Axel to ckeck Olivier's comments on the test cases README.html [recorded in http://www.w3.org/2011/03/29-sparql-minutes.html#action09]

<trackbot> Created ACTION-429 - Ckeck Olivier's comments on the test cases README.html [on Axel Polleres - due 2011-04-05].

LeeF: We have several successful implementations for property paths that path the property path tests
... AndyS, have you run those tests?

AndyS: Yes, there are 30 tests and I have run them

<AndyS> Tests = 30 : Successes = 30 : Errors = 0 : Failures = 0

<LeeF> PROPOSED: Approve the 30 tests in the property-path directory

<AndyS> (error means bad test setup e.g. data syntax wrong ; failure means different results)

<kasei> +1

<AndyS> seconded

<OlivierCorby> +1

<LeeF> RESOLVED: Approve the 30 tests in the property-path directory

<LeeF> ACTION: Lee to mark 30 prop path tests approved today [recorded in http://www.w3.org/2011/03/29-sparql-minutes.html#action10]

<trackbot> Created ACTION-430 - Mark 30 prop path tests approved today [on Lee Feigenbaum - due 2011-04-05].

LeeF: I wanted to look at the negative syntax tests
... bad01 to bad03, they don't have update opertions, AndyS suggests we make that legal, so the tests are no longer negative tests
... can be useful in some applications

<kasei> I'm hesitant, but no strong feelings

LeeF: Paul, Alex, any implications for the upate spec?

Paul (?): I think that wouldn't be a big change.

LeeF: Are you indifferent, in favour, or against that?

Paul (?): I don't have strong feeling, but seems ok

LeeF: I don't see it doing harm and seems easy

AndyS: Changes the grammar
... the grammar is in the query doc even for update

<LeeF> ACTION: Andy to change update grammar to allow zero-operation update requests [recorded in http://www.w3.org/2011/03/29-sparql-minutes.html#action11]

<trackbot> Created ACTION-431 - Change update grammar to allow zero-operation update requests [on Andy Seaborne - due 2011-04-05].

AxelPolleres: AndyS, for INSERT DATA and DELETE DATA, we do have quads there now. Is that intended?

AndyS: That is unrelated to zero operation deletes

AxelPolleres: Yes

AndyS: There is a note that says that there are no variables

<LeeF> PROPOSED: Approve tests in syntax-update-1 except for *bad-0{1,2,3}.ru

AndyS: that's in 19.8

<LeeF> Let's take up all the syntax tests next time

AndyS: I can do the prefix update first and then prprove the tests next week

<LeeF> Adjourned.

<MattPerry> bye

<SteveH> bye everyone

Summary of Action Items

[NEW] ACTION: Andy to change the @prefix : prefix in the syntax directories to use an absolute URI [recorded in http://www.w3.org/2011/03/29-sparql-minutes.html#action06]
[NEW] ACTION: Andy to change update grammar to allow zero-operation update requests [recorded in http://www.w3.org/2011/03/29-sparql-minutes.html#action11]
[NEW] ACTION: Axel to add those new types to mf: and to README.html [recorded in http://www.w3.org/2011/03/29-sparql-minutes.html#action05]
[NEW] ACTION: Axel to ckeck Olivier's comments on the test cases README.html [recorded in http://www.w3.org/2011/03/29-sparql-minutes.html#action09]
[NEW] ACTION: Lee to clean up occurrences of ut:graphData [ ut:data [recorded in http://www.w3.org/2011/03/29-sparql-minutes.html#action07]
[NEW] ACTION: Lee to fix the delete-insert queries that should be negative syntax tests in delete-insert [recorded in http://www.w3.org/2011/03/29-sparql-minutes.html#action08]
[NEW] ACTION: Lee to follow-up and make sure we get IF() tests [recorded in http://www.w3.org/2011/03/29-sparql-minutes.html#action04]
[NEW] ACTION: Lee to mark 30 prop path tests approved today [recorded in http://www.w3.org/2011/03/29-sparql-minutes.html#action10]
[NEW] ACTION: Matt to look at new d-entailment text [recorded in http://www.w3.org/2011/03/29-sparql-minutes.html#action01]
[NEW] ACTION: Matthew to include a test on nodes in path of length zero come from specified named graph (e.g. graph <g1> {?x <p>* ?y}) [recorded in http://www.w3.org/2011/03/29-sparql-minutes.html#action03]
[NEW] ACTION: Matthew to look at new d-entailment text [recorded in http://www.w3.org/2011/03/29-sparql-minutes.html#action02]
 
[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.135 (CVS log)
$Date: 2011/03/29 15:04:25 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.135  of Date: 2009/03/02 03:52:20  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

Succeeded: s/teve/Steve/
Succeeded: s/ut: graph/ut:data/
Found ScribeNick: bglimm
Inferring Scribes: bglimm
Default Present: AxelPolleres, cbuilara, bglimm, LeeF, kasei, OlivierCorby, MattPerry, SteveH__, pgearon, alepas, AndyS, Souri_, NicoM
Present: AxelPolleres cbuilara bglimm LeeF kasei OlivierCorby MattPerry SteveH__ pgearon alepas AndyS Souri_ NicoM
Regrets: Chime NickH sandro
Agenda: http://www.w3.org/2009/sparql/wiki/Agenda-2011-03-29
Found Date: 29 Mar 2011
Guessing minutes URL: http://www.w3.org/2011/03/29-sparql-minutes.html
People with action items: andy axel lee matt matthew

[End of scribe.perl diagnostic output]