See also: IRC log
<kendallclark> http://lists.w3.org/Archives/Public/public-rdf-dawg/2006OctDec/0090.html
Agenda is swell. Let's do it.
<AndyS> Can't anyone can email the list (as Henry demo'ed!)
No minutes public yet from last week, we'll approve them next week
Daylight time changes before next meeting -- for now we'll keep the meeting at 14:30Z
AndyS: not available the week of the 7th (ISWC) but available the 14th
LeeF: availble the 7th but not the 14th (SWEO IG F2F)
PatH: available
EliasT: available both
AndyS: EricP probably not available on the 14th
kendallclark: available both
we'll probably meet both dates even with small crews
Next meeting: October 31, 14:30Z, AndyS to scribe
<scribe> ACTION: FredZ to repost pointers to previous arguments for curlies on OPTIZONAL LHS [DONE] [recorded in http://www.w3.org/2006/10/24-dawg-irc]
<scribe> ACTION: LeeF and EliasT to summarize open SPARQL protocol issues and propose resolutions in email to WG [DONE] [recorded in http://www.w3.org/2006/10/24-dawg-irc]
<scribe> ACTION: PatH to draft replacement vanilla entailment section for WG consideration [CONTINUES] [recorded in http://www.w3.org/2006/10/24-dawg-irc]
<scribe> ACTION: KendallC to close formsOfDistinct issue [CONTINUES] [recorded in http://www.w3.org/2006/10/24-dawg-irc]
<scribe> ACTION: PatH to review the proposed tests in http:// lists.w3.org/Archives/Public/public-rdf-dawg/2006JulSep/0169 and say yay or nay [CONTINUES] [recorded in http://www.w3.org/2006/10/24-dawg-irc]
<ericP> Regards: +EricP
LeeF: consensus seems to be that people feel the need for expressions in the select list, but closing/postponing this issue won't preclude future work by a future group
PROPOSED: to close the puncutationSyntax issue
<kendallclark> a basic scribe need! :)
RESOLVED, FredZ (Oracle) abstaining
kendallclark: last week, the discussion about filter scope shifted quickly from FILTERs to how to determine the first argument to OPTIONAL
<kendallclark> http://lists.w3.org/Archives/Public/public-rdf-dawg/2006JulSep/0243
<kendallclark> Andy's reply:
<kendallclark> http://lists.w3.org/Archives/Public/public-rdf-dawg/2006OctDec/0099.html
-> http://lists.w3.org/Archives/Public/public-rdf-dawg/2006JulSep/0243 contains some test cases about the scope and order of FILTER
AndyS: dawg-filter-scope-003 is
more about the execution order than about the scope of filters
per se
... Lee's design would be consistent with the proposed algebra
change
ericP: can't pass the tests without writing new code because I have variables in scope for the inner group
<kendallclark> cute :)
<ericP> LeeF, i think kendallclark asked you to scribe something about 001 and 004, but i don't know what
<kendallclark> PROPOSAL: to approve dawg-filter-scope-001, 002, dawg-filter-order-001
<AndyS> That better? I'm muted here now.
PROPOSAL: to approve dawg-filter-scope-001, 002, dawg-filter-order-001
<EliasT> yay
RESOLVED
<jeen> yay
<ericP> yay!
<scribe> ACTION: LeeF to throw said 3 approved tests into the approved bucket [recorded in http://www.w3.org/2006/10/24-dawg-irc]
<kendallclark> ACTION: KendallC to the wee, lost filter tests should be put to the question (re: http://lists.w3.org/Archives/Public/public-rdf-dawg/2006OctDec/0099.html) [recorded in http://www.w3.org/2006/10/24-dawg-irc]
<ericP> 4.7: [[@@Filters apply to the whole of the group they are in. Canonically, all matchingis done, then filters are applied. Implementations wil optimize this.]]
<kendallclark> http://lists.w3.org/Archives/Public/public-rdf-dawg/2006OctDec/0080.html
<LeeF> ericP, needs to specifically say the *innermost* group
AndyS: bottom-up algebra approach in http://lists.w3.org/Archives/Public/public-rdf-dawg/2006OctDec/0080.html -- document changes would be rewriting definitions and formally defining join and leftjoin for SPARQL (slightly different from relational algebra)
<ericP> AndyS, do you mind if i patch the @@ to include "innermost"?
LeeF: i think this makes it clearer, addresses some concerns of the community, easier to reason over
SimonR: I think this is a hue improvement
AndyS: Wanted to give anyone who has objections a chance to raise them; it is a significant change, and so I want people to be able to voice concerns before we go too far down this road. In particular, I'd like to hear from Fred
<AndyS> EricP : innermost isn't clear to me. Seems to be the most deeply nested not "same level".
kendallclark: there's a new document from the Chileans trying to work out formal semantics for SPARQL. they claim to formalize the semantics w/o changing the semantics. Would it be worth having this group of people review the changes we are making here?
<ericP> The way we generally handle the ACTION on folks who aren't here is to ACTION some other poor shmuck to pester them
<LeeF> AndyS, "the smallest group which contains the FILTER" -- isn't that accurate ?
<ericP> LeeF, perhaps s/smallest/innermost/ ?
<LeeF> ericP, was trying to avoid innermost since Andy wasn't happy with it :)
<patH> Where are these new Chilean documents? Are they public?
<AndyS> By this point, it's going to probably better be expressed relative to the algebra and syntax words may be confusing.
ericP: how do we communicate in the test materials that a certain test requires certain extensibility in a SPARQL engine?
<SimonR> (Clattery keyboard!)
<kendallclark> path: i'm sending it to the list now
<patH> ta
<LeeF> AndyS, OK I'll be glad to be a word-reviewer rather than a word-smither of this particular wording :)
<AndyS> yes - it's only an @@ for me!
ericP: i proposed mf:requires such that FOO mf:requires xsd:someType means that the test requires support for the type
<kendallclark> (email w/ chilean semantics paper sent to list)
AndyS: harder to talk about tests in the implementation report without having URIs for them
<jeen> (back)
<jeen> zakim unmute me
<AndyS> Woks for me - I'll go and impl new algebra :-)
<SimonR> Maybe giving the tests an rdf:type instead, so that classifying something as in a partcular manifest and required by a particular extension could be done using subclassing.
PROPOSED to adjourn early
ADJOURNED