SPARQL Working Group Teleconference

03 Mar 2009


See also: IRC log


LeeF, kasei, AndyS, +1.216.445.aaaa, iv_an_ru, bijan, [Garlik], Chimezie_Ogbuji, john-l, terraces, AxelPolleres, ivan, +656304aabb, ywang4, +1.603.897.aacc, Souri, EricP, Ori, daveNewman, Souri_Das
ericP, AxelPolleres, Eric P.
minutes from last meeting (Feb 24)


<scribe> scribenick: AxelPolleres

Lee: clarification... Wg is open only to W3C members, although we conduct work in public
... minutes, agenda, wiki are all public and encouraged to be shared.

Lee: if you want to keep something non-public, mail us (team/chairs)


minutes from last time are on the web.

PROPOSED: accept minutes from last week

<bijan> I just look and they look ok

RESOLUTION: minutes from last week approved

next meeting: 10am US time, 1400Z

lastlog trackbot


LeeF: we have intros linked from the wiki
... encourage others to supply intros

john-l: chime and i missed last week

john-l: 2nd rep for cleveland clinic (cleveburg OH)
... chime and i are working on an impl, hoping to push mods back into spec

chime: john-l and i work on a patient registry

chime: we've done a hand full of improvements. expect they are useful in general

ivan: ivan the elder
... W3C semweb activity lead
... 2nd contact

yimin: working in lilly singapore
... working on data integration projects in lilly
... looking for insert/delete, stuff like that

souri: Souri Das, Oracle
... was part of DAWG. passionate about it as i expect it to be useful for DI in general


LeeF, we might want a tracker intro under logistics

<LeeF> ericP, yes, i agree, but am not prepared to do it at the moment :)

AxelPolleres: liasons are supposed to connect us with other groups touching our work
... need liasons with three W3C groups:

<bijan> XQuery?

AxelPolleres: .. OWL
... .. RIF
... .. RDB2RDF

<bijan> HCLS interest group

AxelPolleres: any other suggestions?
... would like liason status update at beginnings of meetings

<bijan> I think it's overkill to do it each telecon...generally, there's fewer occasions to mention

RESOLUTION: Bijan appointed OWL liason

<iv_an_ru> Orri and I are in RDB2RDF

<AxelPolleres> Bijan: additional liaison XQuery?

<bijan> +1 to eric for liasing with HCLS

RESOLUTION: ericP appointed HCLS liason

RESOLUTION: orri appointed RDB2RDF liason

RESOLUTION: axel appointed RIF liason

<Zakim> ericP, you wanted to say i have good contact with HCLS

<AxelPolleres> Should we "advertise" to them? maybe someone was interested to join us, at least to follow?

<AndyS> XQuery are also working on full text search

but i was locally muted

<AndyS> s/arealoo/are also/

ericP: expect that we will need to touch XQuery if we import more XPath functions


LeeF: most folks have tougher travel restrictions than in the past

<iv_an_ru> we will need to touch XQuery if we extend XQuery with SPARQL :)

LeeF: don't expect to have our four meetings this year
... hope to have a f2f a couple months from now when we settle on our deliverables

PROPOSED: host first f2f 5-6May in Cambridge, MA, USA -- (other telecon sites may be amendmants)

<bijan> tues are potentially bad

<AndyS> May 6,7 is preferrable to me as well. Can make May 5 if necessary

<bijan> I can't do tues easily

<AxelPolleres> eric: rooms might be a problem 7-8

<bijan> 2 days

<chimezie> +1 2 days

<ywang4> 2 days should be fine

<SteveH> +1 2 days

<ivan> 2 days (it is hard to keep up work of that intensity for 3 days)

<souri2> May 6-7 is better for me. 2 days.

<AndyS> 2 days unless it is clear 3 is needed

<bijan> Can't do the 5th

<bijan> (most likely)

<AxelPolleres> no preference

<ywang4> okay, i can try to adapt

<iv_an_ru> My US visa has expired so I'm probably missing the f2f.

<AxelPolleres> let's tentatively fix 6-7 shall we also set up a questionaire?

<SteveH> +q to ask about multi site

<ivan> no

<bijan> I think chairs decision is fine

<Zakim> SteveH, you wanted to ask about multi site

<bijan> As long as it's announced far enough in advance...8 weeks

<AndyS> I can ask our comms people about a telecon link

<bijan> I can ask about access grid access

<LeeF> Pick other (U.K. / Europe?) site in the next week or two, please send possibilities to mailing list

PROPOSED: host first f2f 6-7May in Cambridge, MA, USA -- (other telecon sites may be amendmants)

<souri2> What will be the hours for F2F?

<AxelPolleres> http://www.w3.org/2009/sparql/wiki/F2F1

<Zakim> ericP, you wanted to mention video room availability

<bijan> Manchester has a room

<bijan> HP has one...I've seen it :)

use cases

<iv_an_ru> use case 1: TPC-H

<iv_an_ru> (we pass it in SPARQL :)

LeeF: 1st phase: define what features we want to work on
... we're bound to 18 months, including spec, tests, impl report
... likely a very small set. need to settle on them in the next month

<AxelPolleres> Please indicate your availability for F2F1 on the wiki page: http://www.w3.org/2009/sparql/wiki/F2F1

LeeF: we're blurring the line between use case and feature
... e.g. "peform aggregate functions", "query over arbitrary length tasks"
... we're not aiming for higher-level use cases, à la "Bob runs a stereo store..."
... looking for examples, implementations

<iv_an_ru> use case 2: a graph with Gantt chart, personal calendars and some trivial validation.

LeeF: need a champions who will speak in depth about the feature

bijan: would we do one req, or stagger them?

LeeF: should we write todo list or a priority list?

bijan: i like getting things done
... but there's value publicity of a slew of things
... but if we can get stuff out

AndyS: knocking off some easy ones would get the community engaged

<bijan> +1 to AndyS

<SteveH> +1 to AndyS

AndyS: i like to work on one thing at a time, and do it properly

ivan: short charter, and rec process takes time and energy
... publishing several things in a staggered way could backfire
... bijan, is the idea to pub sparql1,1, 1,2, 1.3?

<AxelPolleres> http://www.w3.org/2009/01/sparql-charter.html#deliverables

<iv_an_ru> +q

bijan, was wondering if we could modularize parts into specs.

scribe: e.g. separate owl semantics spec

<bijan> The HTML5 spec indicates the stability levels of different *sections* of the spec...which might be good for implementors

LeeF: modularizing sounds great, if we can modularize ourselves

<AndyS> +1 to modularization (if it does not cause rework just for modulrity)

<Zakim> bijan, you wanted to answer

ericP: could modularize sections, do tests and impl reports, and package as a spec when we get near the end

AxelPolleres: use cases should identify which spec they affect (e.g. protocol, xml results format)
... we've listed deliverables in the charter. we're not bound to that?

LeeF: nope. our deliverables will fall out from our use cases

iv_an_ru: not sure the final doc should be a single spec, e.g. SPARQL1.1
... they [features] are more useful when used together
... when i need business intelligence, i need it on everything

<ivan> +1 to iv_an_ru

LeeF: don't expect us to break up core parts

<Zakim> SteveH, you wanted to ask about terminology

<bijan> The spec shouldn't have everything it but neither should it be the minimal publishable unit

LeeF: don't want to lose interop if all these features are on their own specs

<bijan> Rightsizing!

SteveH: do we want use cases?

<bijan> New features + rationale?

SteveH: am nagged by calling features "use cases"

<iv_an_ru> yes!

SteveH: just want to call them what they are

<chimezie> then perhaps we should be a little more explicit in the request for what we want (usecase v.s. feature request)

<bijan> (The OWL document is called "new features and rationales"

LeeF: agreed. would like to dive straight into features

SteveH: when we prioritize, we'll talk about use cases, regardless of whether we write them down

<AndyS> While features are natural/obvious this is OK but if one comes along which the WG as a whole finds unobvious, we need to back off to use cases.

<LeeF> +1 to AndyS_

<SteveH> +1

AxelPolleres: template rationale: thought the features we want to add should be driven by practical examples

<iv_an_ru> use cases should be written down at least in our wiki, if not in final docs, because they demonstrate how features "interoperate".

<bijan> Use cases doesn't exhaust...e.g., implementability are also considerations

AxelPolleres: had in mind to merge use cases in the 2nd phase
... good feature requests often have a use case around them

<iv_an_ru> Let's think that everything is implementable :)

<bijan> So, new feature: "Awesome feature"; rationale: "here's the intuition", "here's the implementation considerations", "here's some use cases"

AxelPolleres: if you a system with .e.g aggregates, add you syntax to that feature

<bijan> Are we all registered in the wiki? If not what do we do?

<SteveH> +1 to calling the features and having a usecase subsection

LeeF: let's call them features (tweak the template), add a section for use cases under each feature, and have example syntaxes for the feature

<LukeWM> +1 to that too

<ywang4> will go offline shortly, have fun :)

APPROVED: rename "use cases" to "features"

<bijan> Did I miss the in

<bijan> structiosn for getting on the wiki?

<bijan> Oh. Ewwwww

<LeeF> I'll send a bunch about this in email

<Zakim> AndyS, you wanted to ask about outside contributions

AndyS: what about folks who are not in the group, but plan to join
... ?

LeeF: would like folks to send to public-rdf-dawg-comments

<bijan> +1 to open solicitation

<SteveH> risks a lot of work sorting out duplicates

AndyS: danger of opening floodgate

<SteveH> but still +1

<bijan> We can always say "no" :)

<LeeF> I'll take the responsibility for sorting out dupes :-)

<bijan> No specification without championiation!

<iv_an_ru> We can invite people to read the wiki before (re)submitting proposals :)

<LeeF> iv_an_ru +1

<AxelPolleres> heavy echo

<bijan> ta ta

<AxelPolleres> scribenick: AxelPolleres

Ivan: we can allow input from those who intend to join.

Lee: Will sned an email on the list to gather as many features as we can by next week.

<iv_an_ru> Even more, implementation providers may announce in their support mailing lists that they collect wishes for SPARQL 2.0 and act as proxies.

<AndyS> SPARQL 1.1 please. Not 2.0. It's incremental!

Lee: Is it appropriate to list that on the blog as well.

<SteveH> +1 to SPARQL 1.1

Ivan: sure.

<iv_an_ru> ok, SPARQL 1.1

Ivan: Next week will be messy, since meeting is one hour earlier!!!

<SteveH> bye

meeting adjourned

<LeeF> AxelPolleres, do you want to try to follow the scribe instructions, or would you like me to?

Summary of Action Items

[NEW] ACTION: LeeF to see if the SPARQL tracker is setup correctly [recorded in http://www.w3.org/2009/03/03-sparql-minutes.html#action01]
[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.133 (CVS log)
$Date: 2009/03/03 16:26:38 $