See also: IRC log
RESOLVED to accept minutes of 5 Jan as record of previous meeting
RESOLVED: to meet next Thu 12 Jan 1600Z. scribe: Elias
<DanC> you might check your calendar for 17Jan, libby
KC reports progress in http://lists.w3.org/Archives/Public/public-rdf-dawg/2006JanMar/0048.html
<DanC> sparql-protocol-query.wsdl,v 1.15 2006/01/09 19:59:32
<DanC> KC:
kendall: there are now 2
bindings
... WSDL and working groups
understanding are now in sync (?)
<kendallclark> If you want all the gory details, see the checkin message for 1.16
<kendallclark> I was hoping to ask Hugo and someone in DAWG for review.
<patH> hi enrico
<kendallclark> LeeF: http://www.w3.org/2001/sw/DataAccess/proto-wd/sparql-protocol-query.wsdl
kendall: 2 things about the design
that are not obvious: cant have a conformant SPARQL service that
returns N3 (or anything other than RDF/XML for graphs)
... if it accepts an XML query format, thats not legal either
... need to create new binding and interface if you want to do
that
... can say it returns application/*, or list the return type
explicitly
<AndyS> DanC: N3 designers may be after text/*
<kendallclark> sorry, I missed Pat's comments ;>
<kendallclark> I'm happy to list more IMTs explicitly
DanC: I think it should be */*
PatH: agree that its asymmetric (in v's out variation), but dont want multiple input formats
DanC: you have to be prepared to throw away in/out puts that you dont understand
AndyS: input serialisation is about the wrapper, not the query string (?)
<patH> for the record, I'd like multi formats both ways. BUt I think asymmetry with one end freee is betr than both ends constrained.
patH, sorry! misunderstood
<patH> OK
<kendallclark> 2.2.1.11 is an example
[p]
Danc: want to put the WSDL protocol question on thursday
kendall: will add */* to the output serialisations
<AndyS> DanC is correct - I was working through the issues.
<kendallclark> LeeF
thx
LeeF: SOAP is left as RDF/XML root element, with contents unconstrained
<AndyS> One is that we say in rq23 that "A SPARQL query string is a Unicode character string ... in the language defined by the following grammar".
<ericP-telnet> having irc client challenges -- using telnet
LeeF: thats the only way you could return content, eg. put N3 in RDF root element as text
kendall: group of people who want to use SOAP probably dont overlap much with people who want to use N3
<DanC> ACTION: KC to refine .wsdl description w.r.t. mime types [recorded in http://www.w3.org/2006/01/10-dawg-minutes.html#action01]
<DanC> ericp, re-poke re policy foo
<kendallclark> BTW, this ACTION: review sparql-types for xs:any weirdness is finally DONE
<ericP-telnet> will do
<kendallclark> (I say "no work has been done" as the slackerist WG member! :>)
<DanC> ACTION: EliasT to recruit somebody to try out the protocol test harness, perhaps from the -comments list [DONE] [recorded in http://www.w3.org/2006/01/10-dawg-minutes.html#action02]
<DanC> Elias: looks like Richard and I should have something next week
EliasT: just have to update the python tests to be insync with joseki
<ericP-telnet> ACTION: EricP to pester W3C folks about unified comments on security and policy [recorded in http://www.w3.org/2006/01/10-dawg-minutes.html#action03]
EliasT: mostly read automatically from spec, but there are some that are purposly broken (includes ...)
<LeeF> I will.
<LeeF> Elias will point me to the example in question.
<LeeF> I've got an hour to kill writing long, creative SPARQL queries.
<kendallclark> I just don't have time to be more creative! :>
<AndyS> SPARQL allows duplicate triple patterns - it can be :s :p :o x1000
DanC: the exmaple should cross the traditional URL length limit
<DanC> ACTION: Jeen try to reproduces EliasT's protocol testing results [CONTINUES] [recorded in http://www.w3.org/2006/01/10-dawg-minutes.html#action04]
patH: 2nd mail is not sent yet
<DanC> Subject: bnodeification (was: Re: SPARQL semantics: open issues for basic query patterns)
<DanC> Date: Mon, 9 Jan 2006 17:27:19 -0600
patH: mail num. 1 goes beyond the
action item because theres a proposal in there
... allows extra redundancy in the answer set, in order to maintain
the simplicity of the definitions, but there is a way to reduce the
redundancy
<LeeF> EnricoFranconi: latest definitions are not the same as original subgraph matching definitions
EnricoFranconi: it is important for interoperability that services should return same answers from a lean and non-lean graph (?)
<DanC> Pat, Enrico, I need to start wrapping up (unless I hear a proposal to extend)
<DanC> 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/2006/01/10-dawg-minutes.html#action05]
keep going, guys!
patH: almost done with action, but not quite
<DanC> ACTION: Enrico to review draft text on matching and entailment for solution sequences [CONTINUES] [recorded in http://www.w3.org/2006/01/10-dawg-minutes.html#action06]
<Souri> (just fyi) Folks at our org have some comments that we hope to post by end of this week
<DanC> ACTION: Bijan to elaborate on test/example of foaf query where bnodes cross answers [CONTINUES] [recorded in http://www.w3.org/2006/01/10-dawg-minutes.html#action07]
<DanC> (I presume)
<DanC> from Enrico
<DanC> Enrico confirms that's current from him
<EnricoFranconi> http://lists.w3.org/Archives/Public/public-rdf-dawg/2006JanMar/0030.html\
<DanC> Bijan proposes Andy incorporate that
<LeeF_> extend
i need to go
ericP: will take over scribing
<LeeF_> bye, SteveH
<EnricoFranconi> sorry about that
<EnricoFranconi> yes
Enrico: the content of 0030 is roughly correct, but PatH is welcome to tinker with it
Andy: this is clear except for result sets
<DanC> AndyS: I see how to incorproate 0030 , save for results set...
PatH: I can add some explanatory
text
... for result sets
<EnricoFranconi> ok for me
PatH: may I s/RDF merge/ordered merge/ ?
<SerT> sure
<SerT> BTW ordered merge could be used to give a precise definition to the CONSTRUCT operator
<AndyS> Noted
DanC: Andy, the text you incorporate will say "no" to PPS's use case?
AndyS: yes
Bijan: re redundancy on answer sets, does that affect distinct?
<EnricoFranconi> distinct and redundancy are different issues, I believe
<patH> I agree.
Bijan: so DISTINCT operates over syntactic redundancy, not logical redundancy
<DanC> ACTION: Bijan to elaborate on test/example of foaf query where bnodes cross answers [DONE] [recorded in http://www.w3.org/2006/01/10-dawg-minutes.html#action08]
action -1
<DanC> Pat's msg of Mon, 9 Jan 2006 17:27:19 -0600 discharges bijan's action
action -8
<scribe> ACTION: KC to refine .wsdl description w.r.t. mime types [recorded in http://www.w3.org/2006/01/10-dawg-minutes.html#action09]
action -7
<AndyS> ""A test passes if the graph from the action is logically equivalent to the graph named in the result.""
<DanC> ACTION: DanC to review test README re "logically equivalent" vs graph equivalent [recorded in http://www.w3.org/2006/01/10-dawg-minutes.html#action10]
<DanC> ADJOURN.
<patH> quit
<Souri> quit
<EnricoFranconi> bye all!
<LeeF> bye, EnricoFranconi
<SerT> bye everybody
2006-01-07T20: 19:40Z <bjoern_> so if you
have a normative statement that character references must represent
legal XML 1.1 Chars, and an informative note that [^...] excludes
anything that is not a Char per XML 1.1, I'd be happy
... 19:40Z <bjoern_> so if you have a normative statement
that character references must represent legal XML 1.1 Chars, and
an informative note that [^...] excludes anything that is not a
Char per XML 1.1, I'd be happy