Data Access Working Group

30 May 2006

See also: IRC log


Souri, Kendall_Clark, ericP, LeeF, AndyS, FredZ, PatH
LibbyM, SteveH, DanC


PROPOSED: to accept <http://lists.w3.org/Archives/Public/public-rdf-dawg/2006AprJun/0083> as a true record of the 9 May 2006 telecon


1. Convene [1]RDF Data Access WG meeting of Tuesday, May 30, 2006 at

kendall: I was not at XTech or WWW. What community interest was expressed?

<kendall> AndyS: I'd like to hear about Jena community interest too

LeeF: People seem to accept SPARQL as enabling the semweb get around the next corner

<AndyS> Ack

<AndyS> TBL said "SPARQL" in his intro, several times

<kendall> Leigh Dodds gave a talk about SPARQL, including good stuff about protocol

<kendall> EricP presented SPASQL

EricP: Leigh Dodds' and Jim Melton's and ericP's SPARQL talks were well attended

<kendall> 55 people at Jena user conference; lots of interest in SPARQL

<kendall> Ingenta talk about SPARQL & Jena & VLDBs

AndyS: Ingenta's talk at Jena User Conference and XTech came accross positively

kendall: wondering if the JSON works needs more strength behind it


<LeeF> +1 to what Andy said

AndyS: it is in a good state for what most folks need

AndyS: 80% of the folks [somewhere] claimed to know how to write SPARQL queries

2. XTech & WWW SPARQL/DAWG show reports

LeeF: [describes the JSON callback issue]

<AndyS> There is a proposal for a fix for this in JSONRequest so there is a coming solution for all

<LeeF> I've heard much doubt expressed that JSONRequest a) solves the problem and b) will get enough steam to make momentum

<AndyS> http://www.json.org/JSONRequest.html

<kendall> yep, I'm a bit skeptical about JSONRequest.

<AndyS> maybe - but do we want to compete/promote a different solution which has not been security-reviewed? :-)

<kendall> LeeF: can you drop a pointer to this Yahoo doc that describes the callback solution?

LeeF: cross site scripting problem: page a gets access to intranet and gives data to another site.
... cross site scripting demo solution is to shut off browser security.
... [describes yahoo solution]
... problem: can't use POST
... proposal from Doug M-something, who invented JSON, has several security restrictions.

<kendall> Douglas Crockford

LeeF: in my talks, i don't see it gaining critical momentum

<kendall> he didn't invent it as much as wrote it up :)

<AndyS> Doug's slides: http://json.org/json_www2006.ppt including lots of links to stuff of interest

LeeF: Elias, AndyS and I discussed adding reference to the yahoo parameter-named-callback document
... I am not happy with any of these.

<AndyS> Yahoo doc: http://developer.yahoo.com/common/json.html

kendall: agree, none are good
... odd to put in a results document

LeeF: we may have crossed that line when we mentioned mime type

<Zakim> kendall, you wanted to point out MIME type bug in JSON Results format doc

kendall: i don't want to add the callback, but i want to fix the mimetype

<scribe> ACTION: LeeF to fix mimetype in json-reults [recorded in http://www.w3.org/2006/05/30-dawg-minutes.html#action01]

LeeF: DaveB does not want to implement the the callback unless it's in the document

kendall: i don't think we should explain how to get around security restrictions in a results format document

JSON results format -- update on where we stand

valueTesting -- update on where we stand

<LeeF> yes

<LeeF> I'm more comfortable saying nothing if we know that there's active work on the front going on elsewhere

<kendall> EricP: Some browser vendors, including Mozilla, working on the x-site scripting issue, and there are some groups w/in the W3C who are also working on the issue

<kendall> EricP: though some of them may be doing so somewhat informally

<AndyS> Is this issue valueTesting as a whole or just the boolean case?

<kendall> http://www.w3.org/TR/XMLHttpRequest/

<kendall> FWIW

test suite maintenance -- update on where we stand

ACTION: AndyS: to draft of open-world = tests [recorded in http://www.w3.org/2006/05/09-dawg-minutes#action07] [CONTINUED]

<scribe> ACTION: EricP to propose text and tests to add {boolean < > = != boolean} to SPARQL [recorded in http://www.w3.org/2006/05/09-dawg-minutes#action06] [CONTINUED]

Renewed interest in SADDLE?

kendall: some traffic on and off list about saddle
... have folks done enough to work for us to consider doing DAWG work

extraction of the SADDLE input from the Boston F2F

<kendall> AndyS: Issue isn't resolved; believes that we're not able to solve it w/ the resources & time we have left

AndyS: I think it we couldn't do it in the time allotted

<LeeF> +1 Andy and Eric - just not quite ready

DDR Use Case?

kendall: MWI, or someone, has a Device Description Repository
... historically used RDF to store the data. want to use SPARQL to query.
... semweb coordination group wants us to ... maybe use it as a use case?
... Anyone want to look at the DDR doc?

<AndyS> The doc does not contain the string "RDF" - it's use cases

kendall: ericP: I think we can slack on this. They will very likely look at SPARQL and tell us where it hurts

<Souri2> Does SPARQL require that a graph pattern must contain at least one non-OPTIONAL triple-pattern?

<kendall> PROPOSED: to meet on 6 June 2006 at 13:30 UTC

Next meeting: 6 June, chair: KendallC, scribe: ???

<AndyS> Seconded

<kendall> So RESOLVED?

zakime, take up next agendum

Souri's Q:s

<LeeF> For that matter, you can have a WHERE { }

Souri: SELECT ?a WHERE { } -- what's ?a ?

<LeeF> zero rows

<kendall> hmm, FWIW, there's an ISWC proposed paper that claims SPARQL is PSPACE-complete... I bid heavily to review it. :>

<LeeF> or 1 row with an unbound var?

<kendall> but someone here might be interested in following up w/ the authors?

<LeeF> I would rather that these discussions be done on the WG lists.

<EliasT> ericP, thanks for volunteering AndyS w/o his consent.

<EliasT> It's ok to ask questions but to expect the WG to change/reconsider the document after the time has expired for comments, it causes disruption.

<EliasT> However, if you have development questions, we have a public mailing list for development.

<EliasT> http://lists.w3.org/Archives/Public/public-sparql-dev/

<kendall> yes, elias, but these seem to be questions about the spec from a WG member. I don't see any requests for spec changes.

<LeeF> 9:30 boston time

<Souri2> Query => SELECT ?x WHERE {?x a _:y}

<Souri2> Data => <John> a <Human> . <John> a <Mammal> . <John> a <Animal> .

<Souri2> Result = {<John>, <John>, <John>} or just {<John>} ?

<LeeF> Elias will do it.

<EliasT> WHAT?

<EliasT> WHAT?

<EliasT> QUE?

Next meeting: 6 June, chair: KendallC, scribe: EliasT

<kendall> Elias to scribe next week


<EliasT> Souri2, LeeF will answer that question for you.

<EliasT> did Fred just leave?

so it appears

Query => SELECT DISTINCT ?x WHERE {?x a _:y} => {<John>}

<Souri2> Query2: SELECT ?x WHERE {?x a ?y}

<LeeF> sequence of solutions --> projection --> *set* of solutions (represented however an implementation chooses) --> maybe distinct

Summary of Action Items

[NEW] ACTION: EricP to propose text and tests to add {boolean < > = != boolean} to SPARQL [recorded in http://www.w3.org/2006/05/09-dawg-minutes#action06]
[NEW] ACTION: LeeF to fix mimetype in json-reults [recorded in http://www.w3.org/2006/05/30-dawg-minutes.html#action01]
[PENDING] ACTION: AndyS: to draft of open-world = tests [recorded in http://www.w3.org/2006/05/09-dawg-minutes#action07]
[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.127 (CVS log)
$Date: 2006/05/30 16:39:57 $