See also: IRC log
<DanC> sergio gets a gold star for including test cases/examples in his msg of Thu, 5 Jan 2006 12:19:56 +0100
<SteveH> no :(
<DanC> minutes 20 Dec
<DanC> "update" agendum didn't get proper markup, but no biggy
<DanC> RESOLVED to accept DAWG Weekly -- 20 Dec 2005 as a true record
<kendallclark> sorry to be late, my IRC client went bad and I didn't realize it
<DanC> np KC; we're booting
<kendallclark> i also have a bad cough today, so i'll be muted as much as possible
<DanC> http://www.w3.org/2002/09/wbs/35463/exteljan2006/results
<DanC> Jeen Likely to miss january 10th.
<DanC> regrets Janne Saarela 5Jan
<DanC> regrets PatH
<bijan> I'm watching and can switch over at need
<LeeF> Go away, Zakim.
<Jos_De_Roo> Zaki, ??P65 is me
<SteveH> i can do it if i get a mic working by then
<DanC> ok, I'll take it
<DanC> RESOLVED: to meet next 10 Jan, scribe SteveH (we hope)
<kendallclark> Hmm, I had a conflict about scribing next week because I didn't want to do it. Thanks, Zakim!
<kendallclark> rock!
<kendallclark> (i cleared several months old actions!! :>)
<DanC> ACTION: AndyS to respond re CONSTRUCT * [DONE] [recorded in http://www.w3.org/2006/01/05-dawg-irc]
<scribe> ACTION: AndyS to Revise rq23 to reflect the decison (remove told bnode text) [DONE] [recorded in http://www.w3.org/2006/01/05-dawg-irc]
<scribe> ACTION: Jeen to edit results format per C 'collapsed' ; drop sorting constraint. [DONE] [recorded in http://www.w3.org/2006/01/05-dawg-irc]
<scribe> ACTION: Lee to review XSLT in results format update. [recorded in http://www.w3.org/2006/01/05-dawg-irc]
<DanC> Query by Reference
<DanC> KC 4 Jan 2006
Kendall: leigh suggested we allow 'query by reference' in the protocol: point to location for query in request.
... not particularly hard to do, not particularly compelled to add it either, happy to do what the WG wants (correct?)
<kendallclark> (yes)
(jeen has some headphone trouble, difficult to understand speakers)
<AndyS> Hmm
ericP: does not think it's risky, not very motivated to add it either.
<kendallclark> it's strongly analagous to what cwm does, I think. but YMMV.
Kendall: early version of the protocol draft had this feature
<EliasT> I second it.
<kendallclark> bad connect, sorry
<DanC> PROPOSED: to postpone queryByReference , due to lack of detailed implementation experience
<kendallclark> and i don't see any constraint on the size of the postponed issues list :>
<AndyS> +1
<DanC> so RESOLVED.
<DanC> ACTION: KC to review sparql-types for xs:any weirdness, w.r.t. WSDL 2.0 constraints [DONE] [recorded in http://www.w3.org/2006/01/05-dawg-irc]
<DanC> ACTION: KendallC to propose revised WSDL descripton of SPARQL protocol [CONTINUES] [recorded in http://www.w3.org/2006/01/05-dawg-irc]
<kendallclark> we may in fact be WSDL 2.0's earliest adopter :>
<scribe> ACTION: Jeen try to reproduces EliasT's protocol testing results [CONTINUES] [recorded in http://www.w3.org/2006/01/05-dawg-irc]
EliasT: am running it in python and java, using Andy's joseki stuff
<DanC> ACTION: EliasT to recruit somebody to try out the protocol test harness, perhaps from the -comments list [CONTINUES] [recorded in http://www.w3.org/2006/01/05-dawg-irc]
<bijan> One sec
<DanC> from http://www.w3.org/2001/sw/DataAccess/weekly-agenda
<EnricoFranconi> my last thext is at http://lists.w3.org/Archives/Public/public-rdf-dawg/2006JanMar/0018
<DanC> "specific wording proposals 29 Dec and 2 Jan."
<EnricoFranconi> of today
<kendallclark> bad ring on the line now
<DanC> sergion on rdfSemantics with test sketches
<kendallclark> thx eric
<DanC> Re: SPARQL semantics: open issues for basic query patterns
DanC: want to add test case from the example given in http://lists.w3.org/Archives/Public/public-rdf-dawg/2006JanMar/0021.html
<AndyS> The XML Results talks about labels scoped to the document
<bijan> ?x foaf:nick "Bijan". ?x ?p ?o.
<EnricoFranconi> the problem is that the current text is *not* equivalent to the original (and agreed upon) text using matching
<LeeF> sergio's query: SELECT * { ?x ?p ?y }
<LeeF> bijan: More unsettling problem is on coreference within results
<EnricoFranconi> all implementations use graph matching between the query and the data - the current text does not characterise that
<LeeF> patH: criteria should be entailment of the pattern by the graph. if you want proper coreference, include the coreference within the pattern
<LeeF> bijan: I'd expect coreferences between rows in the resultset anyway because if i have 2 people in my FOAF data with foaf:nick "Bijan" -- but other properties differ between the 2 -- it's important to relate the proper properties w/ the proper person
<LeeF> DanC: I don't think any of the proposed designs address this
<LeeF> DanC: none of the current proposed designs talk about anything beyond single query results
<AndyS> Nov 2 doc is Enrico/Sergio's SPARQL semantics doc
<bijan> I'm also going off of this passage in sergio's message:
<kendallclark> i can hear him, but no comprehension
<bijan> """In fact, bnodes in answer sets are really useful only to discover co-
<bijan> references among different pattern solutions (or even within the same
<bijan> one). This means that unless a bnode name is used in the "same way"
<bijan> across different entailment checks, the user cannot trust bnodes in
<bijan> an answer set."""
<DanC> design option, new to me: matching definitions that consider multiple answers, i.e. an answer set. (introduced Nov by Enrico, evidently)
<EnricoFranconi> without coreference bnodes in the answer are useless
<EnricoFranconi> you better get rid of them
<SteveH> my implementation does what bijan said
<AndyS> "Solution Sequence"
<DanC> ... http://www.w3.org/2001/sw/DataAccess/rq23/
<AndyS> sec 10.1
<kendallclark> gone now
<LeeF> heisenecho
<kendallclark> heh
<DanC> "The solution sequence from matching the query pattern is an unordered collection formed from the solutions of the query pattern."
<kendallclark> you hope, dan! :>
<EnricoFranconi> Did anybody read my proposed text?
<LeeF> yeah, that was very sneaky, Dan :)
<AndyS> Yes - I read it.
<EnricoFranconi> comments?
<patH> Enrico, I didnt but I am doing.
<DanC> ACTION: PatH to draft text on matching and entailment for solution sequences [recorded in http://www.w3.org/2006/01/05-dawg-irc]
<AndyS> Please can we discuss the other issue Sergio rasied (the not "unsettling" first one)
<EnricoFranconi> ok
<DanC> ACTION: Enrico to review draft text on matching and entailment for solution sequences [recorded in http://www.w3.org/2006/01/05-dawg-irc]
<EnricoFranconi> but my text is already ok for this :-)
<EnricoFranconi> in fact
<patH> Yes, now I understand answer SET scoping, I think it will be easy to fix all of the issues Sergio raises.
<DanC> Proposed changes by FUB
<LeeF> DanC: Does this proposal need a way to specify entailment types?
<EnricoFranconi> actually I will be happy even by leaving out alternative entailemnts
<EnricoFranconi> FUB is happy to have only the default simple entailemnt
<EnricoFranconi> I'll send a new text today
<kendallclark> indeed
<patH> I think Zakim thinks I am AndyS. Hmm, maybe I am?
<kendallclark> god. worst call quality ever.
AndyS: there are still outstanding comments re rdf merge
<LeeF> good call content, though.
<kendallclark> yes
<LeeF> patH: two orthogonal questions: 1) scoping of bondes within results document 2) do we need to allow alternative entailments to be specified?
<EnricoFranconi> FUB is happy to postpone parametrised entailemnt, but with the current proposed text by FUB today
<LeeF> DanC: 2nd question has a consensus of postponing parametrised entailment, show where it would go in extensions
<EnricoFranconi> point (1) is solved by theorem 1 of the proposed text of today
<EnricoFranconi> so path revises my text I'll send out later today
<patH> OK.
<DanC> ACTION 6 = PatH: revise Enrico's "Proposed changes" on matching and entailment for solution sequences, esp w.r.t. RDFmerge/order
<DanC> jos pls make test case from http://lists.w3.org/Archives/Public/public-rdf-dawg/2006JanMar/0021.html ?
<kendallclark> (oh! jos is here, I didn't realize. Hi Jos!)
<patH> Hi Jos
<DanC> ACTION: JosD to make test case from http://lists.w3.org/Archives/Public/public-rdf-dawg/2006JanMar/0021.html [recorded in http://www.w3.org/2006/01/05-dawg-irc]
<DanC> ACTION: Bijan to elaborate on test/example of foaf query where bnodes cross answers [recorded in http://www.w3.org/2006/01/05-dawg-irc]
<AndyS> There may be one like that already (easy action!)
<bijan> yay!
<bijan> jeen, send them!
<kendallclark> *nasty*
<DanC> ACTION: EricP to add test case for support of xsd:boolean (ala (?a = ?b) = "true"^^xsd:boolean ) [CONTINUES] [recorded in http://www.w3.org/2006/01/05-dawg-irc]
<ericP> [[
<ericP> The canonical representation for boolean is the set of literals {true, false}.
<ericP> ]]
<ericP> http://www.w3.org/TR/xmlschema-2/#boolean
<patH> Will we get grief from i18n for using English words in canonical forms?
<patH> That was NOT a joke.
<patH> Ah, OK.
<AndyS> Should rebuild value from any form. So xfer should be safe.
<Zakim> AndyS, you wanted to ask about Sergio's first example (short?)
<DanC> back to http://lists.w3.org/Archives/Public/public-rdf-dawg/2006JanMar/0021.html
<EnricoFranconi> it is not a typo: it shows that the current text is broken!
<LeeF> Yes
<LeeF> I've got it.
<LeeF> 75 minutes ago
<kendallclark> 75 minutes ago
<LeeF> jinx
<kendallclark> time flies when yr torturing people! :>
<LeeF> No
<kendallclark> 2nd adjournment
<kendallclark> Do that after adjournment?
<DanC> ADJOURN.
<ericP> ACTION: EricP to respond to, or at least consider, http://lists.w3.org/Archives/Public/public-rdf-dawg-comments/2005Dec/0008 [recorded in http://www.w3.org/2006/01/05-dawg-irc]