Minutes of the DAWG telecon of 2005-11-08 for review

RDF Data Access Working Group
http://www.w3.org/2001/sw/DataAccess/
Tuesday 2005-11-08 14:30-16:30 UTC
log: http://www.w3.org/2005/11/08-dawg-irc

chair: DanC
scribe: JanneS

1. Convene, take roll, review agenda

Attending:
     Jeen Broekstra
     Dan Connolly
     Lee Feigenbaum
     Enrico Franconi
     Steve Harris
     Pat Hayes
     Eric Prud'hommeaux
     Jos de Roo
     Janne Saarela
     Andy Seaborne
     Elias Torres

Regrets: KendallC, BijanP, LibbyM

Minutes of RDF DAWG telcon 2005-11-01
http://lists.w3.org/Archives/Public/public-rdf-dawg/2005OctDec/0145.html
ACCEPTED without changes

Next meeting: 15 Nov, scribe: ??

---

2. Next Meeting

Will have one next week 15 Nov

3. Security Considerations

ACTION: EricP to respond to "Security Considerations"
DONE

ACTION: KendallC to add a modified version of 
http://www.w3.org/2001/sw/DataAccess/rq23/#security to protocol editor's 
draft, as well as a pointer to the section in rq23 itself
CONTINUED


4. QL conformance

ACTION: EricP to introduce conformance section and add conformance 
wording where appropriate.


5. toward SPARQL QL updated WD

ACTION Jeen: address 2005-08-30T19:01:08Z from jagannathan.srinivasan

ACTION EricP: mail the WG asking about URIs for functions, re comment 
2005-10-26T22:08:10Z from r.newman 
http://www.w3.org/mid/8CC73085-F6B6-4F56-91B8-8940E55B1648@reading.ac.uk

ACTION DanC: propose changes-from-lc section


6. issues#rdfSemantics

Lengthy discussion whether current LC spec is sufficient or if changes 
are truly needed.

ACTION: Enrico to take a pass through the editor's draft, listing what 
will change with the new semantics understanding ...
DONE

ACTION: PatH to review The Semantics of SPARQL
DONE

---

Adjourned at 16:03 UTC, included the following agenda items with all 
action items as continued, please reply back to these minutes if you 
consider you've done these.

7. issues#wsdlAbstractProtocol

ACTION: DanC to notifty DAWG of WSDL response to our WSDL comments
CONTINUED

ACTION KC: review sparql-types for xs:any weirdness, w.r.t. WSDL 2.0 
constraints
CONTINUED

ACTION: KendallC to propose revised WSDL descripton of SPARQL protocol
CONTINUED

8. Toward CR

ACTION: KendallC to remove uses of "my" in URIs in protocol spec
CONTINUED

9. protocol testing update

ACTION: Jeen try to reproduces EliasT's protocol testing results
CONTINUED

10. test suite maintenance

ACTION: EricP to fix test schema to match manifest with negative tests 
[recorded in
             [32]09/27-dawg-minutes.html#action16]
CONTINUED

ACTION: DanC to follow up re optional test based on op:dateTime triple
CONTINUED

ACTION DaveB: add to test suite the temperature case from comment on 
truth tables in [33]commentor's message
CONTINUED

ACTION ericP: to update sop:isIRI to include the "isuri" synonym
CONTINUED

---

Next meeting: 15-November-2005 at 14:30 UTC.



-- 
Janne Saarela <janne.saarela at profium.com>
Profium, Lars Sonckin kaari 12, 02600 Espoo, Finland
Internet: http://www.profium.com

<DanC> Zakim, next item
<Zakim> agendum 2. "next meeting" taken up [from DanC]
* DanC hopes ericp will call in soon
* DanC wonders if elias would consider chairing
<DanC> Zakim, next item
<Zakim> agendum 3. "Security Considerations" taken up [from DanC]
<DanC> agenda + next meeting, try again
* Zakim notes agendum 11 added
<DanC> Subject: 	[OK?] Re: SPARQL: Security Considerations
<DanC> Date: 	Tue, 8 Nov 2005 08:46:00 -0500  (07:46 CST)
<DanC> ACTION: KendallC to add a modified version of http://www.w3.org/2001/sw/DataAccess/rq23/#security to protocol editor's draft, as well as a pointer to the section in rq23 itself [CONTINUES]
<DanC> Ericp's is done
* RRSAgent records action 1
<DanC> Zakim, next item
<Zakim> agendum 4. "QL Conformance" taken up [from DanC]
<JanneS> http://lists.w3.org/Archives/Public/public-rdf-dawg/2005OctDec/0102.html
<DanC> Zakim, take up item 5
<Zakim> agendum 5. "toward SPARQL QL updated WD" taken up [from DanC]
<LeeF> Chair: DanC
<LeeF> Scribe: JanneS
<DanC> Zakim, take up item 6
<Zakim> agendum 6. "issues#rdfSemantics" taken up [from DanC]
<JanneS> DanC: if you say 'simply entails' it matches with current design, does it?
<JanneS> DanC: bnodes in the query act likes variables (can be bound to different nodes)
<ericP> zakim, please dial ericp=617.2
<Zakim> I am sorry, ericP; I do not know a number for ericp=617.2
<ericP> zakim, please dial ericp-617.2
<Zakim> ok, ericP; the call is being made
<Zakim> +Ericp
<DanC> Zakim, drop ericp
<Zakim> Ericp is being disconnected
<ericP> zakim, please drip ericP
<Zakim> -Ericp
<Zakim> I don't understand 'please drip ericP', ericP
<JanneS> Enrico: alternative to doing skolemization would be current design and the subgraph matching people co
<JanneS> s/co/do/
* DanC wonders if "our latest document" is bound in everybody else's context
<ericP> zakim, please dial ericp-617.2
<Zakim> ok, ericP; the call is being made
<Zakim> +Ericp
<ericP> zakim, please drop ericP
<Zakim> Ericp is being disconnected
<Zakim> -Ericp
<SteveH> Zakim, mute me
<Zakim> SteveH should now be muted
* kendall has joined #dawg
<AndyS_> Home: Input: <a> <b> <c> .  Simple entails _:a1 <b> <c> (and many other triples) so { ?x ?y ?z } gives what?  _:a1 <b> <c>??
<AndyS_> s/Home/Discussion/ 
<AndyS_> http://www.inf.unibz.it/krdb/w3c/sparql/
<franconi> http://www.inf.unibz.it/krdb/w3c/sparql/
<Zakim> +Ericp
<ericP> yehaw!!!
<DanC> -> http://www.inf.unibz.it/krdb/w3c/sparql/ The Semantics of SPARQL 2 Nov 2005
* ericP my mike is more sensitive than i'm used to
<LeeF> :jon :plays :baseball
<LeeF> _:a :plays :baseball
<JanneS> DanC: john plays Baseball, someone plays Baseball -> lean: John plays baseball. Only one answer to (?x ?y ?z)
<JanneS> Enrico: 3 levels: 1. current design where the original graph doesn't include redundancy
<JanneS> 2. no redundancy at all
<JanneS> 3. added redundancy
<ericP> <John> <plays> <baseball>.
<ericP> _:x <plays> <baseball>.
<ericP> _:y <plays> <baseball>.
<ericP> .

<ericP> PatH: a non-lean graph could be a subgraph of a lean graph
<ericP> (seemed important, hope i captured it)
<LeeF> yes, that's accurate.
<ericP> <John> <plays> <baseball>:
<ericP>        <name> "John".
<ericP> _:x <plays> <baseball>;
<ericP>     <name> "Bob".
<ericP> _:y <plays> <baseball>;
<ericP>     <name> "Bob".
<LeeF> patH: current design has blanket assumption that answer bindings must be bound to "things" that already exist in the graph (prevents added redundancy)
<AndyS_> Renaming gets done in the results format so no need here.
<ericP> ?x <plays> <baseball> leans out to:
<ericP> <John> <plays> <baseball>.
<ericP> or
<ericP> <John> <plays> <baseball>.
<ericP> _:z <plays> <baseball>.
<ericP> ?
* jeen has the feeling that this all boils down to what we try to capture with the notion of 'dataset' already
<DanC> Zakim, who's on the phone?
<Zakim> On the phone I see LeeF (muted), EliasT, AndyS_, EnricoF, PatH, DanC, JanneS, SteveH (muted), Jeen, Jos_De_Roo, Ericp
<LeeF> jeen - I thought that was the case except for entailment scenarios that can't be expressed as a single logical closure graph (the OWL-DL little house scenario?).
<LeeF> zakim, unmute me
<Zakim> LeeF should no longer be muted
<patH> Lee- as long as we use 'entails' in the actual statement, we don't have to worry about the logical closure issue.
<LeeF> Right, I understand that patH, thanks.
<EliasT> does no code change, means no ENTAILS keyword?
<AndyS_> Yes
<AndyS_> No ENTAILS keyword
<EliasT> so how does Enrico see us giving people hooks for other entailment levels?
<franconi> I mean with the ENTAILS keywork, which defaults to simple-entailment
<DanC> ack steveh
* Zakim unmutes SteveH
<patH> Theoretical hooks in definitions, not code hooks.
* Zakim sees no one on the speaker queue
<ericP> q+
* Zakim sees ericP on the speaker queue
<SteveH> Zakim, mute me
<Zakim> SteveH should now be muted
<ericP> q-
* Zakim sees no one on the speaker queue
* ericP now understands DanC's ack was to unmute steveh
<JanneS> JanneS: lc ok, hopes the editorial wording would match with a more generic semantics roadmap -> no need to change existing code
<EliasT> I'm still confused, patH says theoretical hooks, Enrico says ENTAILS keyword. which one is it?
<franconi> both!
<EliasT> :-)
<ericP> Stick with LC:
<ericP> I want people to fold the semantics parameter into either the service graph or the FROM/FROM NAMED graphs.
<ericP> I haven't seen interop use cases motivating pushing this parameter into the protocol or the grammar.
<ericP> Peter may have problems running test cases that rely on non-lean graphs if his engine requires test cases. Do we have any?
<JanneS> Jos: ok with LC
<patH> Enrico, do you mean that the query will specify which entailment to use? 
<EliasT> if entails defaults to our current design, then we don't have to change the test cases, but code would have to change to throw an error if a different entail parameter is passed.
<AndyS_> algebra has not been worked through and I fear problems will arise if we rush believing it is trivial (e.g. undistinguished variables, unbounds)
* SerT has joined #dawg
<DanC> q+ to clarify that owl entailment can't be "hidden" in the dataset
* Zakim sees DanC on the speaker queue
<patH> Andy, yes. We need to do the checks.
<LeeF> thanks, DanC, that's whay I was trying to say in IRC.
<franconi> Pat, there are several possibilities here: for example theh service may offer different addresses for the different entailments
<JanneS> I thought I have voted for having the different entailments being provided at different addresses...
* ericP zakim, please mute ericP
* Zakim Ericp should now be muted
<DanC> 3 columns http://www.w3.org/2001/sw/DataAccess/issues#rdfSemantics
<EliasT> what does different address mean? different endpoints or different default-graph-uri? both?
<kendall> hmm, while not officially here, i need to think about multiple endpoints in the WSDL, whether there's any issue with having multiple endpoints for the same (from the WSDL point of view) operation
<LeeF> What are the current objections to (1) specifying the LC design w/ entailment verbiage and (2) omitting ENTAILS, hence requiring that any other entailment be encoded in an endpoint or graph URI?
<kendall> ah, good question elias
<JanneS> (I thought different endpoints, Elias)
<DanC> Zakim, take up item 5
<Zakim> agendum 5. "toward SPARQL QL updated WD" taken up [from DanC]
<patH> Lee, thats the gaol Im trying to steer us to.
<AndyS_> OWL disjunction : distinguished / non-distinguished variables change the answer
<patH> gaol/goal (!)
<AndyS_> ie. chnage the SELECT and the change 
<EliasT> I guess we've moved on to next agendum.
<ericP> ack ericP
* Zakim unmutes Ericp
<AndyS_> ie. chnage the SELECT and the # solutions change 
* Zakim sees DanC on the speaker queue
<LeeF> AndyS_, I'm sorry - which example is that, again, that illustrates the distinguished/non-distinguished variable issue?
<AndyS_> Enricos worker example or Bijans (will get links...)
<LeeF> I know that example, no need for the link. Thanks.
<DanC> DONE: ACTION: Enrico to take a pass through the editor's draft, listing what will change with the new semantics understanding ...
<LeeF> I just didn't remember it involving the # of bindings changing with the projection changing. Will re-look at it.
<DanC> DONE: # ACTION: PatH to review The Semantics of SPARQL
<DanC> ACTION: PatH to sketch tests that characterize impact of semantics work
* RRSAgent records action 2
<DanC> -- continues
<AndyS_> Bijan: http://lists.w3.org/Archives/Public/public-rdf-dawg/2005JulSep/0498.html
<DanC> ACTION Bijan: estimate impact of "abstract syntax entailment" etc. on WG test harness [CONTINUES]
* RRSAgent records action 3
<AndyS_> Enricos ==> http://lists.w3.org/Archives/Public/public-rdf-dawg/2004JulSep/0069
<DanC> ACTION: Bijan to take a pass through the editor's draft, listing what will change with the new semantics understanding, and what will not [CONTINUES]
* RRSAgent records action 4
* ericP suffering from convenient turrett's syndrome
* DanC "the paper"?
<DanC> ACTION: Enrico extend http://www.inf.unibz.it/krdb/w3c/sparql/ to cover "the algebra", i.e. optional, UNBOUND
* RRSAgent records action 5
<DanC> Zakim, close item 6
<Zakim> agendum 6, issues#rdfSemantics, closed
<Zakim> I see 7 items remaining on the agenda; the next one is
<Zakim> 4. QL Conformance [from DanC]
<franconi> Semantics of optional: http://lists.w3.org/Archives/Public/public-rdf-dawg/2005JulSep/0451
<ericP> http://www.w3.org/TR/xquery/#id-xquery-conformance
<AndyS_> Enrico - there was a Q from Steve - did it get an answer?
<franconi> Which Q?
<AndyS_> In email - see the lists
* ericP goes hog wild on the spec
<AndyS_> Enrico - see http://lists.w3.org/Archives/Public/public-rdf-dawg/2005JulSep/0457.html
<DanC> EricP is convinced a conformance section is worthwhile; should be editorial diff to rq23
<franconi> Ah, I lost it :-) Oh yes, this is it - I agree
<DanC> DanC suggests getting the relevant protocol text before going to the commentor; WG decision not needed.
<DanC> Zakim, close this item
<Zakim> agendum 5 closed
<Zakim> I see 6 items remaining on the agenda; the next one is
<Zakim> 4. QL Conformance [from DanC]
<DanC> Zakim, close item 4
<Zakim> agendum 4, QL Conformance, closed
<Zakim> I see 5 items remaining on the agenda; the next one is
<Zakim> 7. issues#wsdlAbstractProtocol [from DanC]
<DanC> Zakim, take up item 5
<Zakim> agendum 5. "toward SPARQL QL updated WD" taken up [from DanC]
<DanC> -> http://lists.w3.org/Archives/Public/public-rdf-dawg/2005OctDec/0176.html handling SPARQL QL comments
<DanC> -> http://www.w3.org/mid/97C9F8341808C244BFF39DCB6EDD5AD71116B2@server.home.ryan.levering.name # 2005-08-17T13:55:31Z from RRLevering
* kendall has quit IRC (Ping timeout)
<AndyS_> http://lists.w3.org/Archives/Public/public-rdf-dawg-comments/2005Aug/0096
<DanC> 2005-08-30T19:01:08Z from jagannathan.srinivasan
<DanC> ACTION Jeen: address 2005-08-30T19:01:08Z from jagannathan.srinivasan
* RRSAgent records action 6
* JosD have to leave tcon now.. (having holiday)
<Zakim> -Jos_De_Roo
* LeeF has left #dawg
* LeeF has joined #dawg
<LeeF> whoops.
* LeeF removes ESC key from keyboard
<ericP> { ?x sparql:isIRI ?y }
<DanC> ... FILTER ( my_renamed_sop:isIRI(?x) )
<ericP> vs.
<ericP> ... FILTER ( isIRI(?x) )
<SteveH> no opinion
<ericP> AndyS: i prefer that all functions have a URI
<ericP> ... including the level-breakers
<SteveH> s
<ericP> i.e.  FILTER ( isBound(?x) )  vs.  FILTER ( my_renamed_sop:isBOUND(?x) )  vs.  
<DanC> ACTION EricP: mail the WG asking about URIs for functions, re comment 2005-10-26T22:08:10Z from r.newman http://www.w3.org/mid/8CC73085-F6B6-4F56-91B8-8940E55B1648@reading.ac.uk
* RRSAgent records action 7
<JanneS> DanC suggests publishing of new WD with Eric's and Andy's changes + approval from Mr. X by 16th
<JanneS> EricP: could take the time to issue new LC
<DanC> ACTION DanC: propose changes-from-lc section
<JanneS> DanC: who could do section-by-section change log?
* RRSAgent records action 8
<EliasT> will that publication have answers to all of the comments in http://lists.w3.org/Archives/Public/public-rdf-dawg/2005OctDec/0176.html ?
<DanC> (I'm somewhat inclined to not decide today on the grounds of time)
<AndyS_> Some are susbtantive which is one place where we need WG members help
<AndyS_> i.e. take an issue - propose changes (diffs to HTML best) - send to WG.
<AndyS_> (the right answer is "current text" unless you find a bug :-)
<EliasT> we don't have to have specific assignments.. right?
<EliasT> everyone shold be looking at them and answering what they can... concurrency is solved by email reply to the comment, correct?
<DanC> ADJOURN.
<AndyS_> No - just pick one - tell WG if it will be long thought process
<DanC> RRSAgent, make logs world-access
<RRSAgent> I have made the request, DanC
<DanC> RRSAgent, pointer?
<RRSAgent> See http://www.w3.org/2005/11/08-dawg-irc#T16-03-18