10 Mar 2009


See also: IRC log


Phil Brooks




<LeeF> good job, zakim

<LeeF> Scribe: Phil Brooks

<AxelPolleres> is phil there? john Clark would be the next on the scribe list (http://www.w3.org/2009/sparql/wiki/Scribe_List)

<LeeF> AxelPolleres, did you just send another mail about the time change? I haven't received one

<LeeF> john-l, can you scribe today if Phil isn't around?

<john-l> LeeF: Yes, I can.

<LeeF> john-l, thanks

<john-l> Zakim: aabb is me

<SteveH> Zakim: ??PP32 is me

<kjetil> Zakim: what is the code?

<LeeF> ack ??P41

<ywang4> sorry for a bit late - Yimin

<john-l> Zakim: please mute me

<LeeF> Scribenick: Phil


<LeeF> PROPOSED: Approve minutes at http://www.w3.org/2009/03/03-sparql-minutes.html

<LeeF> RESOLVED: Approve minutes at http://www.w3.org/2009/03/03-sparql-minutes.html

Minutes from last week approved.

next meeting

Next Meeting: 10:00am EST on March 17

Face to Face

<LeeF> Face to Face wiki page

<ywang4> it seems i could not login the wiki with my w3c account...

<AxelPolleres> +1 to have standard setup for "normal" dial-in as well.

<LeeF> ywang4, I'll ask the Wiki Folks about it

<ywang4> thanks

F2F scheduled from May 6-7. In person attendance encouraged. Everyone should check their calendars and let Lee know who will be attending (for planning purposes).

Cost at MIT is $20/day. Looking for sponsors for the MIT site.

Parties interested in sponsoring the MIT site should contact Eric.

(cheap advertising!) :-)

<AxelPolleres> ... two options for alternate location: Bristol and Manchester.

<AndyS> Bristol

Opportunity for 2nd F2F at TPAC2009 in Santa Clara, CA the week of Nov. 2-6, 2009. What does everyone think?

Ivan: Santa Clara may be too expensive. W3C considering alternate location - will decide in several weeks.

<iv_an_ru> For me, CA in Nov would be nice contrast with Novosibirsk in Nov :)

ISWC is week before TPAC2009; attending both could be problematic/costly

ISWC is Oct. 25-29

<AxelPolleres> co-location with ISWC would be an option.

<ericP> ISWC2009 is 25-29 October per http://iswc2009.semanticweb.org/

Alternately, F2F could occur at ISWC

TPAC may be a better option as there will be opportunity to integrate with other W3C groups.

<LeeF> feature prpodsals

<ericP> i wonder if we could use chad for rating the priorities

Lee and Axel's role is to ensure that the "right" list of features is adopted (i.e. Lee wants to make everyone unhappy - at least, a little!)

overview of features

Lee reminds that the group is responsible for creating the specifications for the features that are adopted (and that is not trivial).

<AxelPolleres> Point 1: We won't probably be able to do all in 18month, that might guide us to narrow down the number.

<AxelPolleres> Support by implementations important, pervasive support, usefulness. Can requested features be done with the language as is?

<AndyS> Point 3: SPARQL already has some extension points

<AxelPolleres> option: standardize on extensibility points rather than extensions themselves.

<AndyS> Framework for features to converge outside of this WG process

The features that get adopted must be interoperable (and should be core to the language). SPARQL v1 already has some extensibility points; just because we "could" standardize an extension does not mean we "have to". It may be a better use of our time to standardize on extensibility points rather than to standardize an extension in and of itself. For example, standardizing a service option may advance SPARQL further (especially with respect to interoperabilit

It's unrealistic to review all the proposed features on a teleconference; participation on the SPARQL working group must include work outside of the call. In particular, please start threads on the mailing list regarding your "favorite" feature and why. Others should reply to the posts, ask questions, clarify as needed, challenge as desired, etc. In summary, the mailing lists (and associated threads) will be the main method for discussing the various feature

<Zakim> ericP, you wanted to suggest a mailing convention like Subject: [AccessingRdfLists] blah blah blah

<AxelPolleres> +1 to distinct mail-subjects

<iv_an_ru> There are two features: BI and Federation. Others are minor details of these two :)

<chimezie> There definately seems to be some overlap

<AxelPolleres> +1 to offline go over the list and try to propose merge to authors.

There may be some features that could be merged - Lee encourages anyone interested in doing so to review which are candidates for merging and then make such a proposal.

All work group participants come from varying backgrounds and with different opinions - we will try to reach consensus. However, we all need to maintain an open mind as we review the features and consider which ones to go forward with.

<iv_an_ru> First we should agree about some objectives that are more precise than "better SPARQL, upward-compatible"

<AxelPolleres> SteveH, ok that overconsolidation should be avoided, but try to make links *and differences* with related features explicit.

<ericP> as of last night, my top priorities were update and lists



<AndyS> I would relate it to constructors for RDF terms.

<chimezie> i think it returns a boolean

<chimezie> pure post-processing

Ivan asked clarifying question about first example - the "(?age > 18)" projects a boolean value

<SteveH> sorry, it's a wacky first example

<SteveH> mea culpa

<john-l> No, I think it's a fine first example, once you understand what's going on.

<LeeF> AxelPolleres: Might be one or the other with assignment

<ericP> i agree, and assignment is more powerful

<iv_an_ru> +q

Syntax debates should be reserved for the mailing list

<john-l> Perhaps say "whether the person is over 18" instead of "where the person is over 18"?

<Souri> +1

<john-l> +0

<SteveH> +1

<ericP> +0

Polling of opinions regarding Project Expressions (poll on IRC)

<AlexPassant> +1


<ivan> +1

<kjetil> +0

<kasei> +1

<AxelPolleres> +1 (modulo redundancy avoidance w.r.t. assignment)

<ywang4> +1

<LukeWM> +1

<iv_an_ru> +1

<chimezie> +1

<AndyS> +1

<kasei> and agree with AndyS regarding the potential for rewriting

<LeeF> +1

In general, the group seems to be strongly in favor of Project Expressions.



Next topic: subselects

<LeeF> SteveH: subselects gives you other features "for free"

<LeeF> ... e.g. !ASK for some forms of negation

In addition to Virtuoso as listed on the Wiki, there are other implementers of sub-select as well.

<Zakim> ericP, you wanted to say point out that unlike SQL, we've not needed subselects so far for e.g. coherent OPTIONALS (outer joins) and UNIONS.

(at least 2, anyway)

<chimezie> +q to suggest we consider restricted subselects rather than all or nothing

<AndyS> The cost to the algebra is actually quite small to allow nesting queries. One issue: breaks the multiset/sequence distinction but it's fixable and probably needed any way

<ivan> ?supp+>tpcd:acctbal


<AndyS> which are a sub-ASK

<Souri> What does "+>" in ?supp+>tpcd:acctbal stand for?

<AxelPolleres> Is +> doable with bracketed syntax?

<SteveH> AxelPolleres: apprently because the example after is suposed to be equivalent

Simpler examples on sub-select wiki might help. There have been simpler social-networking examples in the mailing lists.

<SteveH> sorry, AxelPolleres, not :

<john-l> I'm not familiar with TPC.

<chimezie> i am not familiar, unfortunately

<LukeWM> I'm not either

Many folks are not familiar with the TPC queries; therefore, the simplest queries should be used for all of the feature categories (instead of the more complex TPC queries).

<Zakim> chimezie, you wanted to suggest we consider restricted subselects rather than all or nothing

<Zakim> SteveH, you wanted to disagree with eric

<LeeF> SteveH: adding too many surface syntax features makes query language complicated

<Souri> I am somewhat familiar, but we should decouple the examples from TPC benchmark to explain the utility.

<SteveH> +1

<AlexPassant> 0

<john-l> +0

Polling for sub-selects:

<kasei> +0

<LukeWM> +1

<iv_an_ru> +1

<kjetil> +1

<chimezie> +0

<ericP> SteveH, i understand your point. would like to look in detail with you

<ericP> -1

<Souri> +0.75

<ivan> +0

<LeeF> Orri: 1


<LeeF> David: +1

<LeeF> +0

<AxelPolleres> +0

<AndyS> +1

<SteveH> ericP, yes, sure

<ericP> my hunch is that the verb count ends up the same

<chimezie> take care

meeting adjourned.

<ericP> i gotta run now, or i'd erpsue

<SteveH> ericP, I have some "work" to do, but catch me in the future

<LeeF> Phil, thanks for scribing, I will take care of the minutes from here

Summary of Action Items

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.133 (CVS log)
$Date: 2009/03/10 15:02:53 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.133  of Date: 2008/01/18 18:48:51  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

Succeeded: s/Brussels(?)/Bristol/
Succeeded: s/braketted/bracketed/
Found Scribe: Phil Brooks
Found ScribeNick: Phil

WARNING: No "Present: ... " found!
Possibly Present: AlexPassant AndyS AndyS_ AxelPolleres Chimezie_Ogbuji DaveNewman David Ivan LeeF Lee_Feigenbaum LukeWM Orri P29 P32 P41 P48 P53 P59 PROPOSED Phil Scribenick Souri SteveH aabb aacc aaff aagg aahh chimezie ericP iv_an_ru john-l kasei kjetil option sandro trackbot ywang4
You can indicate people for the Present list like this:
        <dbooth> Present: dbooth jonathan mary
        <dbooth> Present+ amy

Regrets: Bijan

WARNING: No meeting title found!
You should specify the meeting title like this:
<dbooth> Meeting: Weekly Baking Club Meeting

Agenda: http://www.w3.org/2009/sparql/wiki/Agenda-2009-03-10
Got date from IRC log name: 10 Mar 2009
Guessing minutes URL: http://www.w3.org/2009/03/10-sparql-minutes.html
People with action items: 

WARNING: Input appears to use implicit continuation lines.
You may need the "-implicitContinuations" option.

[End of scribe.perl diagnostic output]