DAWG Weekly

Convene, take roll, review records and agenda

 

 

<DanC> zakim seems lagged

<AndyS> Sergio, saw your message: the fix to the second part is to scope G' to the whole query, right? No more is needed.

<ericP> http://www.w3.org/2006/01/10-dawg-minutes

<AndyS> The first part (distinguished variables) : I realised we don't really define the scope of variables and could handle it there. e.g.

<LeeF> didn't think you would

<DanC> Scribe: Elias

<EliasT> Scribe: EliasT

<AndyS> If SELECT'ed, then distinguished. If mentioned in two separate BGPs in teh same query ,distinguished (inc bnodes). Else not distingusihed.

<kendallclark> (not $$ charity! :>)

<SerT> Andy, it's difficult to scope G' to the whole query because the algebra is separated form the BGP

<SerT> u need to use the same graph every time

<AndyS> So same G' for the whole query execution?

<SerT> yes

<patH> Yes, neds to be same, but does not need to be G. See my email reply just sent.

<kendallclark> yeah, birth certificate if I can find it. ;>

<DanC> http://www.w3.org/2006/01/10-dawg-minutes

<AndyS> Sergio, and about part 1 - that is about scoping of vars - can we explicitly define what we need?

<DanC> AndyS at risk for end of 17Jan meeting

<DanC> scribe 17 Jan: AndyS, backed by DanC

<DanC> scribe 17 Jan: AndyS, backed by DanC, KC

<DanC> RESOLVED to meet next 17 Jan, noting DanC at risk

<AndyS> Sergio - are you dialing into the telecon?

<DanC> Jos is also offering to scribe 17 Jan.

<DanC> though at risk

<SerT> I'm trying but I've got problems with the skype account

<DanC> whee!

Issue: wsdlAbstractProtocol

<DanC> ACTION: KC to refine .wsdl description w.r.t. mime types [DONE] [recorded in http://www.w3.org/2006/01/12-dawg-minutes.html#action01]

<AndyS> http://www.w3.org/2001/sw/DataAccess/proto-wd/sparql-protocol-query.wsdl

<DanC> sparql-protocol-types.xsd,v 1.17 2006/01/11 19:15:22

<DanC> sparql-protocol-query.wsdl,v 1.17 2006/01/11 19:14:45

<DanC> revision 1.17

<DanC> date: 2006/01/11 19:14:45; author: kclark; state: Exp; lines: +4 -4

<DanC> changed input and output serializations according to telcon on Tue (added "application/xml" to POST's input

<DanC> and */* to GET's output

<DanC> ----------------------------

<DanC> revision 1.16

<DanC> date: 2006/01/09 20:14:21; author: kclark; state: Exp; lines: +7 -5

<DanC> - added faultSerialization="*/*" to queryHttpGet and queryHttpPost

<DanC> ----------------------------

<DanC> revision 1.15

<DanC> date: 2006/01/09 19:59:32; author: kclark; state: Exp; lines: +52 -43

<DanC> - updated WSDL namespace IRIs

<DanC> - Singin', come out if things aren't said, Shoot an apple off my head

<DanC> - you can't bind the same operation twice with the same name, so I split the HTTP binding of SparqlQuery.query into

<DanC> a queryHttpGet binding and a queryHttpPost binding. Which kinda sucks, but works.

<DanC> - removed wsdlx:safety="true" for the query operation, since we give it a POST binding (it's still safe in the relevant

<DanC> sense, but it's not safe at the HTTP level)

<DanC> - I either had binding faults wrong or the WSDL changed, but I replaced <fault name="MalformedQuery"...> with <fault

<DanC> ref="tns:MalformedQuery"...>, since binding fault components don't have a name property, but rather a ref property

<DanC> - (fault|input|output)Serialization attributes now work roughly like Accept: headers in HTTP... You can specify more than

<DanC> one IMT or you can specify IMT ranges (though WSDL discourages using range specifiers). Accordingly, I:

<DanC> - added whttp:outputSerialization="application/sparql-results+xml, application/rdf+xml" to queryHttpGet binding

<DanC> (this means if you want to return results in another IMT, you make a new binding, minimally)

<DanC> - replaced queryHttpPost's whttp:inputSerialization type (which was, inexplicably, set to "multipart/form-data") with

<DanC> the proper value "application/x-www-form-urlencoded"

<DanC> - set queryHttpPost's outputSerialization to the same as queryHttpGet's

<DanC> whttp:outputSerialization="application/sparql-results+xml, application/rdf+xml, */*"

<patH> I have to run away for about 3 minutes, will stay dialled in and muted.

<DanC> PROPOSED: that sparql-protocol-query.wsdl,v 1.17 , sparql-protocol-types.xsd,v 1.17 proto-wd/Overview.html 1.93 addresses issue wsdlAbstractProtocol

<DanC> roger, patH

<ericP> http://www.w3.org/2001/sw/DataAccess/tests/data/i18n/kanji-01-results.ttl to XML

<patH> back now.

<AndyS> What XML can one send in POST? Add application/xml+rdf for same hook for RDF?

AndyS: can there be both parameters in the URL and XML in the content body

LeeF: thinks so

<DanC> (I presume andy means "... an match this WSDL interface decl")

<AndyS> Yes - we need to serialize the request and query.

kendallclark: It's perfectly legal for an endpoint implementation to change how to take the in messages whether url parameters or content body in the HTTP implementation.
... our WSDL only specifies bindings but says nothing how the service tag needs to be defined.

<AndyS> An answer of "it's OK" is fine.

kendallclark: we could say that any changes in the service to our bindings could be considered illegal SPARQL, but it won't be illegal WSDL.
... WSDL tools will try their best to do what we have designed.

<kendallclark> blah blah blah!

kendallclark: blah blah blah.

<kendallclark> heh :>

<kendallclark> no one does, as far as i know

<DanC> PROPOSED: that sparql-protocol-query.wsdl,v 1.17 , sparql-protocol-types.xsd,v 1.17 proto-wd/Overview.html 1.93 addresses issue wsdlAbstractProtocol

<DanC> PROPOSED: that sparql-protocol-query.wsdl,v 1.17 , sparql-protocol-types.xsd,v 1.17 proto-wd/Overview.html 1.93 section 2. SPARQL Protocol addresses issue wsdlAbstractProtocol

<AndyS> Editorial : "Section 5: References: normative" but it has B informative : s/5/5.A/ ??

<DanC> PROPOSED: that sparql-protocol-query.wsdl,v 1.17 , sparql-protocol-types.xsd,v 1.17 proto-wd/Overview.html 1.93 section 2. SPARQL Protocol (less 2.2.1 )addresses issue wsdlAbstractProtocol

<kendallclark> yep, that's right Andy

<DanC> RESOLVED, PatH abstaining

<DanC> (marsh is really pending. http://lists.w3.org/Archives/Public/public-rdf-dawg-comments/2005Dec/0046 )

<DanC> pls visit http://www.w3.org/2001/sw/DataAccess/lc-status-report.html

<DanC> HTTP Status Codes for QueryRequestRefused

<DanC> * 2005-11-01T19:40:55Z from leigh

<DanC> * 2006-01-05T05:01:42Z from kendall

<DanC> ACTION: KC to send an [OK?] msg re HTTP Status Codes for QueryRequestRefused and to paul etc. [recorded in http://www.w3.org/2006/01/12-dawg-minutes.html#action02]

Toward updated protocol WD (and results format WD)

<DanC> ACTION: EricP to propose text on behalf of the W3C team about security and policy [recorded in http://www.w3.org/2006/01/12-dawg-minutes.html#action03]

ericP: He has a spoken comment from Thomas that he would like to put on text

<DanC> a WSDL 2.0 validator

kendallclark: I'm a satisfied with the security and policy text in the protocol draft. However, Graham has almost convinced that there are some issues that need to be addressed.
... It would make sense in the W3C folks could harmonize and give me one text, but if not I'll have to do it myself.
... If ericP sends me a list of the URIs with the issues, I'll address all of the issues this afternoon.

<DanC> ACTION: EricP send KC all the team input on policy, cc DanC [recorded in http://www.w3.org/2006/01/12-dawg-minutes.html#action04]

<DanC> ACTION: KC consider policy input and let us know what text results [recorded in http://www.w3.org/2006/01/12-dawg-minutes.html#action05]

<DanC> action -3

<DanC> . ACTION: Jeen try to reproduces EliasT's protocol testing results

<DanC> WITHDRAWN.

<kendallclark> DanC: May I remove the shaming list of pending comments on the doc?

<jeen> http://www.w3.org/2001/sw/DataAccess/tests/data/rf1/

<DanC> yes, KC

<DanC> mf:result <rf1-example-mixedbindings.srx>

<DanC> JB: steve is aware

jeen: I have added another test for ordered/unordered results. It can be found at http://www.w3.org/2001/sw/DataAccess/tests/data/rf1/

DanC: I was unaware of support for .srx files in QL test suite.

<DanC> PROPOSED: to publish protocol editor's draft version 1.89 2006/01/07 21:31:23 + @@changes reviewed by @@whom, results format version 1.78 of 2006/01/04 15:59:22 + @@changes (example of... darn; forgot) reviewed by @@whom as 2nd Last Call WD.

<DanC> PROPOSED: to publish protocol editor's draft version 1.93 policy section changes reviewed by KC, EP/DC; results format version 1.78 of 2006/01/04 15:59:22 as 2nd Last Call WD.

<DanC> "A conformant SPARQL Protocol service"

<DanC> hypertext anchor, pls

<DanC> KC: ok

kendallclark: I have added a new term that defines "A conformat SPARQL Protocol service".

<DanC> "This set of documents comprises the ^specification of^ the SPARQL Protocol"

<kendallclark> *grr*

<libby> in my calendar and everything

<kendallclark> coolio

<DanC> PROPOSED: to publish protocol editor's draft version 1.93 policy section changes reviewed by KC, EP/DC + hypertext anchor in conformance; results format version 1.78 of 2006/01/04 15:59:22 as 2nd Last Call WD, with comments due 27 Jan (provided we can publish near 13 Jan)

<libby> is it now?

<DanC> yes, started an hour ago. join the fun.

<LeeF> 15 minutes left today, libby

<DanC> 15 more min

<libby> sorry again

<DanC> (hmm.. I see 1.95 now...)

<kendallclark> A conformant SPARQL Protocol service may provide a default RDF dataset against which SPARQL query requests are executed in cases where there is no RDF dataset specified in the protocol or in the query request. A conformant SPARQL Protocol service may refuse to process any query request that does not specify an RDF dataset. Finally, a conformant SPARQL Protocol service may refuse to process any query request against any specified RDF dataset. See 2.1.4 query Fau

<DanC> 1.94 and 1.95 are TOC tweaks, evidently

<kendallclark> yep

<DanC> PROPOSED: to publish protocol editor's draft version 1.95 + policy section changes reviewed by KC, EP/DC + hypertext anchor in conformance; results format version 1.78 of 2006/01/04 15:59:22 as 2nd Last Call WD, with comments due 27 Jan (provided we can publish near 13 Jan)

<kendallclark> Ah, there is an HTML anchor for "conformant SPARQL Protocol service":

<kendallclark> http://www.w3.org/2001/sw/DataAccess/proto-wd/#conf

<kendallclark> not a very good one, mind

<DanC> so RESOLVED.

<DanC> ACTION: DanC to noodle on status [recorded in http://www.w3.org/2006/01/12-dawg-minutes.html#action06]

<DanC> ACTION: KC to tell EricP when you're done [recorded in http://www.w3.org/2006/01/12-dawg-minutes.html#action07]

<DanC> ACTION: EricP to publish [recorded in http://www.w3.org/2006/01/12-dawg-minutes.html#action08]

<SerT> I cannot dialin, system says the conference's restricted

<libby> try *0 sert

<libby> ask amy toput you through

<LeeF> kendallclark: One outstanding comment from Reto

<DanC> http://www.w3.org/2001/sw/DataAccess/lc-status-report.html

issues rdfSemantics, owlDisjunction

<kendallclark> Ah, well, Reto says his comments are "minor, editorial" so maybe we're okay anyway

issues#valueTesting

<ericP> EBV result is the same as xsd:boolean

<ericP> false^^xsd:boolean = "0"^^xsd:boolean

<ericP> true^^xsd:boolean = "1"^^xsd:boolean

<ericP> T=T T=1 1=T 1=1 F=F F=0 0=F 0=0

<LeeF> ACTION: PatH revise Enrico's "Proposed changes" on matching and entailment for solution sequences, esp w.r.t. RDFmerge/order [CONTINUES] [recorded in http://www.w3.org/2006/01/12-dawg-minutes.html#action09]

<LeeF> ACTION: Enrico to review draft text on matching and entailment for solution sequences [CONTINUES] [recorded in http://www.w3.org/2006/01/12-dawg-minutes.html#action10]

<LeeF> # ACTION: JosD to make test case from Sergio's basic query patterns examples [CONTINUES]

<LeeF> ACTION: JosD to make test case from Sergio's basic query patterns examples [recorded in http://www.w3.org/2006/01/12-dawg-minutes.html#action11]

<LeeF> ACTION: DanC to review test README re "logically equivalent" vs graph equivalent [CONTINUES] [recorded in http://www.w3.org/2006/01/12-dawg-minutes.html#action12]

<Zakim> DanC, you wanted to note bug: rdfs:value should be rdf:value

<LeeF> Jos_De_Roo: rdfs: value -- the full URI -- is nonexistent

<LeeF> (above the Data)

<AndyS> boolean-logical-OR

<LeeF> NotApproved

<LeeF> Various lexical forms of xsd:boolean equal an EBV..

<LeeF> Data

<DanC> "Various lexical forms of xsd:boolean equal an EBV.."

<LeeF> ericP: already fixed double period typo

<AndyS> New test is boolean-EBV-canonical ??

<DanC> ADJOURN.