SPARQL Working Group Teleconference -- 28 Jul 2009

<LeeF> trackbot, start meeting

<trackbot> Date: 28 July 2009

<kasei> ericP: ah, ok. wiki lists you as scribing today, but I was asked last week... preference?

<AxelPolleres> http://www.w3.org/2009/sparql/wiki/Vacation_List

<AlexPassant> :)

<AxelPolleres> http://www.w3.org/2009/sparql/wiki/Agenda-2009-07-21

<AxelPolleres> just tarting, birte

<kasei> who is scribing?

<AxelPolleres> http://www.w3.org/2009/sparql/meeting/2009-07-14

<LukeWM> ericP was on scribe list I think

<kasei> I was asked, but not listed on the wiki...

<AxelPolleres> http://www.w3.org/2009/sparql/wiki/Agenda-2009-07-28

<AxelPolleres> scribe: Eric

<AxelPolleres> scribenick ericP

<AxelPolleres> http://www.w3.org/2009/sparql/wiki/Agenda-2009-07-21

<AndyS> Agenda points to 2009-07-14

<AxelPolleres> PROPOSED: accept minutes http://www.w3.org/2009/sparql/meeting/2009-07-21

<AxelPolleres> RESOLVED: accept minutes http://www.w3.org/2009/sparql/meeting/2009-07-21

<kasei> sure

<AxelPolleres> greg to scribe next week.

Liasons

ericP: nothing to report from HCLS or XQuery

AxelPolleres: nothing to report from RIF

<AxelPolleres> no news from liaisons

actions

<AxelPolleres> http://www.w3.org/2009/sparql/track/actions/16

<AxelPolleres> continued

<AxelPolleres> http://www.w3.org/2009/sparql/track/actions/19 continued

<SimonKJ> I have started to get 403 errors on the wiki, it worked 10 minutes ago

<AxelPolleres> http://www.w3.org/2009/sparql/track/actions/55

<SimonKJ> all

<SimonKJ> I had loaded the agenda first thing, I just hot reload and got 403

<SimonKJ> hopefully a temporary glitch, just be nice to see the action/issue list

<AxelPolleres> http://www.w3.org/2009/sparql/wiki/Vacation_List

<AxelPolleres> http://www.w3.org/2009/sparql/track/actions/61 done

<pgearon> +q

<AxelPolleres> http://www.w3.org/2009/sparql/track/actions/62 done

<LukeWM> http://www.w3.org/2009/sparql/wiki/UpdateIssues

<AxelPolleres> http://www.w3.org/2009/sparql/track/actions/63 done

<AxelPolleres> http://www.w3.org/2009/sparql/track/actions/68 done

Establish task-forces for some time-permitting features

<AxelPolleres> action-55 continued!

<Zakim> kjetil_, you wanted to propose that F&R should be early in the meeting

<AxelPolleres> we will discuss F&R before negation.

LeeF: should we be using issue tracker or the wiki?

AxelPolleres: goal of the page is to aggregate e.g. update issues on the update page

<AndyS> Is the Update page supposed to have all the things w have discussed so far?

<Zakim> LeeF, you wanted to ask about http://www.w3.org/2009/sparql/wiki/UpdateIssues vis a vis http://www.w3.org/2009/sparql/track/issues

<Zakim> kjetil_, you wanted to say that HTTP RESTful is missing from the agenda too

<LeeF> ACTION: Lee to create (& link) tracked issues for everything listed at http://www.w3.org/2009/sparql/wiki/UpdateIssues [recorded in http://www.w3.org/2009/07/28-sparql-minutes.html#action01]

<trackbot> Created ACTION-70 - Create (& link) tracked issues for everything listed at http://www.w3.org/2009/sparql/wiki/UpdateIssues [on Lee Feigenbaum - due 2009-08-04].

kjetil: would like HTTP ReSTful update on the agenda today

AxelPolleres: wanted to start with issues we've already discussed. happy to get to [HTTP ReSTful update]

kjetil, update is high-level discussion -- would like that before detailed discussion

kjetil: update is high-level discussion -- would like that before detailed discussion
... would like a straw poll

<LeeF> can someone explain to me what the high-level discussion is that's different from what we'd already resolved?

AxelPolleres: will put ReSTful update on next agenda

<kjetil_> LeeF, have we resolved anything regarding HTTP REST Update?

<LeeF> kjetil, yes, we resolved to pursue them wherever they make sense

AxelPolleres: while we need to spend telecon time on required features, we don't want to forget time-permitting
... would like to find task forces which will meet at other times
... need one or two folks to drive these issues

<LeeF> kjetil, back on June 2 - see http://www.w3.org/2009/sparql/meeting/2009-06-02#resolution_3

AxelPolleres: would like poll progress in a month

<LukeWM> AndyS, I think that the update issues discussed so far ought to be on the wikipage, but no one has added anything apart from me as far as I can tell.

<bglimm> I am volunteering for other entailment regimes

<AndyS> I see a lot of things :: "Negation" and "Time Permitting Features"

LukeWM: update wiki page doesn't cover all issues -- encourage group to flush it out

<kjetil_> LeeF, we're sidetracking here, but after that, there has come up four concrete alternatives to pursue, some of which are very different

AxelPolleres: think that's covered with LeeF's action to transfer issues list issues to update wiki page

<LeeF> kjetil, ok, that's what i was asking about, since i've been out of the loop for a couple of weeks

<kjetil_> LeeF, for example, I and SimonS have forwarded diametrically different proposals

<AxelPolleres> BGP extensions for entailment regimes

<AxelPolleres> http://www.w3.org/2009/sparql/wiki/Feature:SPARQL/OWL

AxelPolleres: for BGP entailment regimes, had 5 folks interested
... Birte and Bijan available to lead Entailment Regime TF
... expect a report in a month to provide info for F&R doc

<AxelPolleres> LeeF: suggests strawpoll per feature.

<AndyS> I am interested in BGP entailment regimes (to implement some of them)

<AxelPolleres> strawpoll: who wants to contribute to BGP entailment TF?

<bglimm> +1

<ivan> I am interested by the BGP one

<AxelPolleres> (+1 means yes, -1 no)

<AxelPolleres> +1

<ivan> +1

<kjetil_> -1

<LukeWM> -1

<SimonS> -1

<kasei> -1

<AndyS> +1

<SimonKJ> -1

-ei^2

<pgearon> +1

<AxelPolleres> Feature: PropertyPaths

<AxelPolleres> http://www.w3.org/2009/sparql/wiki/Feature:PropertyPaths

<AxelPolleres> AndyS takes lead on PropPaths

AndyS: i will drive the PropertyPaths TF

kjetil: I expect to provide test cases for PropertyPaths

<LukeWM> +1

<AlexPassant> +1

<SimonKJ> +1

<ivan> +1

<LeeF> +1

<pgearon> +1

<kjetil_> +1

<AxelPolleres> +0

<SimonS> +0

<AxelPolleres> Feature: Commonly used SPARQL functions

LeeF: it would be great if the TFs meet

ericP: TFs can "/invite Zakim #sparql" and "Zakim, space for 5?" to get an ad-hoc conference

<AxelPolleres> Eric: feature function library I could provide input there.

LeeF: this one will be easier so doesn't need so much formalism

AxelPolleres: would like someone responsible

<AxelPolleres> Axel: I will kick off function library discussion

<AxelPolleres> ericP: existing functions and operators table provides a good way to extend the core.

<AxelPolleres> AndyS: rather identifying URIs for common functions.

s/extensibility mechanism provides a good way to extend the core.

2009-07-28T14: 36:02Z <AxelPolleres> AndyS: rather identifying URIs for

ERG!

<AxelPolleres> strawpoll: who to contribute in function library?

<AxelPolleres> +1

<AxelPolleres> http://www.w3.org/2009/sparql/wiki/Feature:BasicFederatedQuery

<AxelPolleres> feature: federated query

AxelPolleres: champions AndyS, iv_an_ru, SimonS
... TF lead: SimonS

<kjetil_> +1 (hopes to contribute test cases again)

<pgearon> +!

<SimonKJ> +1

<pgearon> +1

<kasei> +1

<LeeF> I hope to contribute, but not sure I'll have enough time

<AndyS> as for LeeF

AxelPolleres: we have folks assigned to all features, please go ahead and start your task forces

minus vs. unsaid

F&R document

kjetil: F&R FPWD published
... my time is uncertain
... have inserted tempates for time-permitting features
... need to get them into shape

AlexPassant: will have time in Aug

AxelPolleres: need to get ahold of time-permitting TF leads to get paragraphs
... would like next F&R at same time as FPWDs of Query and Update

ericP: would be nice to give time to get public feedback, but not sure that's feasible given the timeline

AxelPolleres: do we need more advertising?

<kjetil_> freetext has been brought up

kjetil: swig?

<kjetil_> has it been posted to the semantic-web ml?

<kjetil_> it has been chumped on #swig

AxelPolleres: please suggest publication locations
... we shoudl contact liased WGs

<AxelPolleres> ACTION: Axel to re-check where to announce F&R (e.g. liaisons) [recorded in http://www.w3.org/2009/07/28-sparql-minutes.html#action02]

<trackbot> Created ACTION-71 - Re-check where to announce F&R (e.g. liaisons) [on Axel Polleres - due 2009-08-04].

AndyS: when's last F&R, last call [for Query and Update]?

ivan: it's a note, so we decide the process

kjetil: would like to make a wiki page copying the F&R doc so folks can edit it

<AxelPolleres> ACTION: Alex to take F&R current version back to wiki for editing [recorded in http://www.w3.org/2009/07/28-sparql-minutes.html#action03]

<trackbot> Created ACTION-72 - Take F&R current version back to wiki for editing [on Alexandre Passant - due 2009-08-04].

MINUS vs. UNSAID

LeeF: AndyS written up UNSAID and ericP's written MINUS

<kjetil_> probably a good idea to create new pages rather than use the Category:Features pages

LeeF: from implementor's perspectice, MINUS is an algebraic op to combine result sets
... while UNSAID is a FILTER which takes a triple pattern and examines existing bindings
... from query writer's perspective, in most cases MINUS and UNSAID give you the same results
... when they don't it's when OPTIONALs introduce vars
... those queries tend to be strage and artificial

<AxelPolleres> BTW: I will assign actions to the TF leads to kick-off discussions, just to keep track.

LeeF: someone listed a use case, examining where it is more natural to express one way or another

LukeWM: if we banned opts, would it make MINUS and UNSAID equivalent?

LeeF: i think we think they're identical at that point

LukeWM: just wondering if there are compelling use cases for OPTIONALs

<kasei> i'm not sure it can be identical if we haven't nailed down the antijoin(+restriction) issue that ericP was discussing on the list yesterday.

<Zakim> AxelPolleres, you wanted to comment on OPTIONALs

LeeF: proposal to prohibit OPTIONALs in the unsaid pattern is interesting

<LeeF> kasei, you're right, but i've been working on the assumption that we are dealing with AntiJoin+Restriction, which is the same semantics as UNSAID for the cases in which the two patterns share no variables I think

AxelPolleres: i think restricting patterns makes it harder

<pgearon> +q

<AndyS> Messy syntax but that's all, I think.

pgearon: leaning towards eliminating OPTIONALs because you don't want to eliminiate solutions which would have !BOUND vars

<SimonS> +1 pgearon.

<AxelPolleres> if we disallow that, then we probably should equally think about forbidding nesting optionals... ?!?

LeeF: who's uncomfortable with prohibiting OPTIONALs in MINUS

<AxelPolleres> +1 to eric's +1 ;-)

<kasei> for the unsaid version, I think it makes a lot of sense, since I believe optional is a no-op in that case. not sure about minus.

<AndyS> Unsure about this because optionals appear in the LHS so the compatibility rules seem odd.

<SimonKJ> Could go either way on this, there's complexity in both choices

ericP: would rather the discomfort occur when folks have to understand what happens with unbound vars rather than have different rules for subtraction sets

<AxelPolleres> LeeF: as long as the semantics is clear, for the corner cases, I think we don't have to worry too much about corner cases arising from OPTIONAL use

<AxelPolleres> (hope that summarizes lee right)

Leef: MINUS (modulo ignoring empty results), UNSAID, OPTIONAL prohibited

<pgearon> we *could* say that OPTIONAL on the right is undefined. Is that a terrible idea if there is no use case to use it?

<LeeF> MIUS-AntiJoin+Restriction

<LeeF> MINUS-AntiJoin+Restriction

<LukeWM> +1

<AndyS> -1

<SimonKJ> -1

<SimonS> -1

<ivan> 0

<AxelPolleres> 0

<LeeF> +1

<kjetil_> 0

<AlexPassant> 0

<bglimm> 0

<kasei> i don't want to drag things down here, but i'm still shaky on the equivalence of unsaid and the antijoin+restriction...

<Prateek> 0

<kasei> -1

<kasei> maybe I need to re-read the emails

<LeeF> UNSAID

why do they have to be equiv?

<AndyS> +1

<kasei> +1

<AlexPassant> +1

<LukeWM> -1

<ivan> +1

<AxelPolleres> +1

<SimonKJ> +1

<SimonS> +1

<LeeF> +1

<Prateek> +1

<kasei> ericP: they don't but I thought that had been asserted for the case without optionals (perhaps I've misuderstood that, too, though :)

<pgearon> abstain on both

<pgearon> probably for similar reasons to LeeF

<LeeF> MINUS or UNSAID prohibiting OPTIONALs on the RHS

-1

<kjetil_> 0

<kasei> +1

<AndyS> -1

<SimonS> 0

<SimonKJ> 0

<LukeWM> +1

<AxelPolleres> 0 (not entirely egainst it, since it seems it might drop worst case complexity)

<ivan> 0 (does not know yet...)

<Prateek> 0

<LeeF> -1

<bglimm> 0

<kasei> although I think this might be two different things depending on UNSAID or MINUS

<pgearon> I'd rather "undefine" this pattern, rather than prohibit it

<LukeWM> AndyS, sure

LeeF: clear leaning towards UNSAID

<AxelPolleres> LeeF/Eric: at least in FPWD we should sketch both.

LeeF: consistent with having AndyS edit the template
... plus red spec text inclusing alternatives

wrap-up

AxelPolleres: i'll make actions for TF leads
... action to get TF leads to kick off discussion

<kasei> AndyS: is your understanding that optional is a no-op in UNSAID?

<LeeF> eric, can you try to follow the minutes instructions? :)

- DRAFT -

SPARQL Working Group Teleconference

28 Jul 2009

Attendees

Contents

Liasons

actions

Establish task-forces for some time-permitting features

minus vs. unsaid

F&R document

MINUS vs. UNSAID

wrap-up

Summary of Action Items

Scribe.perl diagnostic output