Chatlog 2011-04-05

See original RRSAgent log and preview nicely formatted version.

Please justify/explain all edits to this page, in your "edit summary" text.

13:56:01 <RRSAgent> RRSAgent has joined #sparql
13:56:01 <RRSAgent> logging to http://www.w3.org/2011/04/05-sparql-irc
13:56:01 <LeeF> trackbot, start meeting
13:56:03 <trackbot> RRSAgent, make logs world
13:56:05 <trackbot> Zakim, this will be 77277
13:56:05 <Zakim> ok, trackbot; I see SW_(SPARQL)10:00AM scheduled to start in 4 minutes
13:56:06 <trackbot> Meeting: SPARQL Working Group Teleconference
13:56:06 <trackbot> Date: 05 April 2011
13:56:14 <LeeF> trackbot, close ACTION-353
13:56:14 <trackbot> ACTION-353 Update the comments table on the wiki and approach people for drafting responses closed
13:56:19 <LeeF> zakim, this will be SPARQL
13:56:19 <Zakim> ok, LeeF; I see SW_(SPARQL)10:00AM scheduled to start in 4 minutes
13:56:48 <Zakim> SW_(SPARQL)10:00AM has now started
13:56:54 <Zakim> +??P2
13:57:01 <SteveH> Zakim, ??P2 is me
13:57:01 <Zakim> +SteveH; got it
13:57:03 <Zakim> +??P3
13:57:26 <AndyS> zakim, ??P3 is me
13:57:26 <Zakim> +AndyS; got it
13:57:34 <AndyS> zakim, who is on the phone?
13:57:34 <Zakim> On the phone I see SteveH, AndyS
13:57:44 <MattPerry> MattPerry has joined #sparql
13:58:40 <Zakim> + +1.310.729.aaaa
13:58:45 <kasei> Zakim, aaaa is me
13:58:45 <Zakim> +kasei; got it
13:58:46 <Zakim> +LeeF
13:59:53 <Zakim> +MattPerry
14:02:30 <bglimm> bglimm has joined #sparql
14:02:33 <kasei> scribenick: kasei
14:03:06 <LeeF> zakim, who's on the phone?
14:03:06 <Zakim> +Sandro
14:03:08 <Zakim> On the phone I see SteveH, AndyS, kasei, LeeF, MattPerry, Sandro
14:03:41 <Zakim> + +41.86.528.aabb
14:03:51 <LeeF> topic: Admin
14:03:55 <LeeF> PROPOSED: Approve minutes at http://www.w3.org/2009/sparql/meeting/2011-03-29 
14:04:01 <bglimm> Zakim, +41.86.528.aabb is me
14:04:02 <Zakim> +bglimm; got it
14:04:30 <LeeF> RESOLVED: Approve minutes at http://www.w3.org/2009/sparql/meeting/2011-03-29 
14:04:40 <bglimm> Zakim, mute me
14:04:40 <Zakim> bglimm should now be muted
14:04:49 <LeeF> Regrets: Chime, Axel, Alex, Carlos, Olivier
14:05:02 <LeeF> Next regular meeting: 2011-04-12 @ 15:00 UK / 10:00 EST (scribe: Axel or Alex) 
14:05:24 <Zakim> +pgearon
14:05:27 <AndyS> My regrets for next week.
14:05:27 <Zakim> +Souri_
14:05:34 <Souri> Souri has joined #sparql
14:05:37 <SteveH> I'm at risk for next week
14:05:57 <kasei> LeeF: new comments. one from last friday.
14:06:22 <sandro> scribe: sandro
14:06:23 <kasei> ... about response codes for protocol. is 204 acceptable in case where there is no content returned?
14:09:19 <AndyS> 204 means empty entity -- we do return an entity for an empty results table in SELECT.  No "semantic" level meaning.  it also means "no change of focus." unlike 200
14:09:57 <sandro> zakim, agenda?
14:09:57 <Zakim> I see nothing on the agenda
14:10:08 <sandro> sandro has changed the topic to: agenda: http://www.w3.org/2009/sparql/wiki/Agenda-2011-04-05
14:10:21 <LeeF> Agenda: http://www.w3.org/2009/sparql/wiki/Agenda-2011-04-05
14:10:31 <sandro> SteveH: There's a bit of the Algebra I still need to look at.       Very short on time, though.
14:11:03 <sandro> Lee: Steve, there are some Aggregrate-related comments for you, too...  Maybe you can look at them at the same time.
14:12:04 <sandro> AndyS: Lee, are you still going to do your action?
14:12:12 <sandro> Lee: I'm not sure, but I hope to.
14:12:44 <LeeF> close ACTION-409
14:12:45 <trackbot> ACTION-409 Edit Update document to note that blank nodes in DELETE templates are an error closed
14:12:46 <sandro> Lee: Paul, has Update Doc been updated to reflect blank nodes being prohibited in templates?
14:12:53 <sandro> Paul: Yes.
14:13:50 <sandro> Lee: Greg, did we say we were ready for you to review update?
14:14:01 <sandro> Greg: We did, I hope to do it this week.
14:14:17 <sandro> Lee: I hope to do a large amt of work on Protocol Doc today.  
14:14:44 <sandro> Lee: I have an email to Chime about come outstanding comments on Graph Store Protocol doc.
14:14:45 <bglimm> Zakim, unmute me
14:14:45 <Zakim> bglimm should no longer be muted
14:14:54 <sandro> Lee: We got a review from Jeff Pan, which Birte replied to.
14:15:30 <sandro> bglimm: I need some clarification from Jeff, but I don't think there's anything that will need WG discussion.   But we do have to discuss issues with Canonical Representation for D-Entailment.
14:15:50 <sandro> Lee: Fed Query still waiting on Carlo or me to check status.
14:16:01 <sandro> q+ to give an announcement
14:16:14 <LeeF> ack sandro
14:16:14 <Zakim> sandro, you wanted to give an announcement
14:16:34 <LeeF> sandro: New working group - provenance working group just announced
14:16:39 <LeeF> ... first meeting in a couple of weeks
14:16:51 <LeeF> ... sign up!
14:17:02 <LeeF> ... also government linked data working group has been sent to the Advisory Committee
14:17:25 <sandro> http://www.w3.org/2011/gld/charter.html
14:17:33 <sandro> http://www.w3.org/2011/prov/wiki/Main_Page
14:18:25 <LeeF> sandro: RDFa WG has been renamed RDF Web Applications WG, working on RDF Web Api
14:18:31 <LeeF> ... (re-)starting up, good time to join
14:18:47 <LeeF> topic: Canonicalization
14:19:20 <sandro> Lee: I didn't read enough of the discussion between Birte and Bijan on this; can you summarize?
14:20:23 <sandro> bglimm: XSD 1.0 canonical reprs have some problems.   Decimal and integer have a bit of a contradiction.   Take canon of the base type.   Which alas means for int it's like 1.0, even though that's not a valid int rep!
14:20:55 <sandro> bglimm: the next version is scheduled for LC in April, this month.
14:21:05 <sandro> LeeF: If they really are on that schedule.
14:22:39 <sandro> sandro: (tells story of OWL and RIF xsd finesse note.)
14:22:47 <sandro> Lee: Should we do the same thing, Birte?
14:23:08 <sandro> bglimm: Yes, I think so.  Anything else -- using XSD 1.0 -- doesn't make sense.
14:23:27 <sandro> Lee: It doesnt seem to help anyone to point at the defns in  a broken spec.
14:23:49 <sandro> Lee: I kinda think we oughta do the same thing as OWL 2 and RIF, depending on 1.1
14:24:05 <sandro> Lee: ... if they don't actually move forward faster than us.
14:24:18 <sandro> Lee: I think that's what we have to do.  :-|
14:24:36 <sandro> bglimm: That sounds right to me.
14:24:48 <sandro> Lee: I don't think we need a WG decision on this.
14:25:12 <sandro> AndyS: At the worst, we could lift a def'n from 1.1 just before we go to CR.
14:25:19 <LeeF> topic: Service URL vs. URI vs. Endpoint
14:25:24 <bglimm> Zakim, mute me
14:25:24 <Zakim> bglimm should now be muted
14:25:41 <LeeF> http://lists.w3.org/Archives/Public/public-rdf-dawg/2011AprJun/0000.html
14:25:45 <sandro> Lee: Terminology from a comment by Bob Scanlon, on svg descr doc.
14:26:50 <sandro> Lee: What is the name of the URI that services SPARQL protocol requests.     "Service URL", "Service Name", "SPARQL Ends Point" (the URI at which a SPARQL protocol service listens....)
14:27:02 <sandro> Lee: Greg suggested "service URI"
14:27:30 <sandro> AndyS: i think the naming has to reflect the difference between the service and where you access it.
14:27:54 <sandro> AndyS: Using "URI" suggests the service ...    I think we need to use the word "End Point"
14:27:58 <sandro> +1 End Point
14:28:10 <LeeF> http://www.w3.org/2009/sparql/docs/service-description-1.1/xmlspec.xml#sd-url
14:28:43 <LeeF> Andy is suggesting that sd:url relates the service to the service endpoint
14:28:50 <sandro> LeeF, : sd:url relates the service to the service endpoint
14:29:19 <sandro> q+
14:29:24 <sandro> q-
14:29:34 <LeeF> SERVICE <foo>
14:29:40 <sandro> q+
14:29:47 <LeeF> kasei: foo is an endpoint
14:29:59 <LeeF> ack sandro
14:31:09 <LeeF> sandro: is the range of sd:url a URL or a character string?
14:31:25 <LeeF> ... is it a literal or not?
14:31:34 <sandro> sandro: In my mind, a URL is a kind of Literal.
14:31:36 <LeeF> kasei: it's meant to point at the endpoint, not point at the literal with a URL string
14:32:15 <sandro> sandro: sd:endPoint or endPointAddress (if it's a string)
14:32:58 <sandro> endpoint 
14:33:09 <sandro> +1 sd:endpoint
14:33:32 <sandro> AndyS: i wouldn't use a capital in the middle :-)
14:33:55 <sandro> Lee: I think it makes sense to call the property "endpoint"
14:34:22 <sandro> kasei: Without concern for people who might be using this already, yeah, I have a mild preference for endpoint, too.
14:34:30 <LeeF> service-identifier sd:endpoint service-endpoint-url
14:35:17 <sandro> Lee: where  service-endpoint-url is a non-Literal Resource.
14:35:47 <sandro> (mild digression about whether that could a blank node.)
14:37:24 <sandro> AndyS: the endpoint is necessarily 1-1 to the endpoint URL
14:37:33 <sandro> sandro: what about changes in case of the hostname?
14:38:02 <sandro> AndyS: It's still 1-1 --- different case in hostname is a DIFFERENT ENDPOINT -- because you'll have a different base for your queries and responses.
14:38:17 <LeeF> ACTION: Greg to make sure SD terminology is consistent with "service endpoint" as the place you poke to use a (protocol) service
14:38:17 <trackbot> Created ACTION-432 - Make sure SD terminology is consistent with "service endpoint" as the place you poke to use a (protocol) service [on Gregory Williams - due 2011-04-12].
14:38:31 <LeeF> ACTION: Lee to work with Carlos to make sure Fed terminology is consistent with "service endpoint" as the place you poke to use a (protocol) service
14:38:32 <trackbot> Created ACTION-433 - Work with Carlos to make sure Fed terminology is consistent with "service endpoint" as the place you poke to use a (protocol) service [on Lee Feigenbaum - due 2011-04-12].
14:38:40 <sandro> sandro: then, yeah, I guess i'm okay with endpoint's range being EndPoint, instead of URL.
14:39:14 <sandro> Lee: let's change sd:url to sd:endpoint -- it'll help communitcate our view
14:39:26 <sandro> sandro: sd:endpoint range sd:Endpoint  ?
14:39:39 <sandro> kasei: That's not the sort of thing we currently do.
14:39:57 <LeeF> topic: RDF dataset merge
14:40:06 <kasei> sandro, I'll think over that...
14:40:44 <sandro> Lee: RDF-WG discussion - what does it mean to merge datasets?
14:41:08 <sandro> AndyS: putting the defns side by side
14:41:22 <sandro> LeeF: Would that change any semantics, or is this just editorial?
14:41:37 <sandro> AndyS: I wouldn't use the word merging.
14:41:50 <sandro> AndyS: Union affects graph store.
14:42:04 <sandro> AndyS: ... around handling of bnodes, in a way that doesnt apply in query.
14:42:37 <sandro> LeeF: My main concern was that since our specs dont use Merge, we'd be including text we dont need, that's not right for RDF-WG.
14:43:08 <sandro> AndyS: Normally I'd agree, but I can't think of another defn of RDF Merge that would work with the definitions we have.
14:43:55 <sandro> Lee: I'm hearing that we use RDF datasets, and given how we define them, there's really only one way to merge them.  And then RDF-WG can use SPARQL Datasets, or something else if they don't like our merge.
14:44:46 <sandro> sandro: Maybe ask RDF-WG to look at this at their F2F next week.
14:45:10 <sandro> Lee: Andy, can you put it in SPARQL Query, and I'll raise it to RDF-WG?
14:45:30 <LeeF> ACTION: Andy to add RDF Dataset merge definition to Query document
14:45:31 <trackbot> Created ACTION-434 - Add RDF Dataset merge definition to Query document [on Andy Seaborne - due 2011-04-12].
14:45:40 <LeeF> ACTION: Lee to ping RDF WG about RDF Dataset Merge definition once in place
14:45:40 <trackbot> Created ACTION-435 - Ping RDF WG about RDF Dataset Merge definition once in place [on Lee Feigenbaum - due 2011-04-12].
14:45:52 <LeeF> topic: tests
14:45:58 <LeeF> subtopic: property path tests
14:46:29 <sandro> LeeF: a bug was found....
14:46:42 <LeeF> http://www.w3.org/2009/sparql/docs/tests/data-sparql11/property-path/
14:46:57 <LeeF> :pp29 in http://www.w3.org/2009/sparql/docs/tests/data-sparql11/property-path/manifest.ttl
14:47:07 <LeeF> path-3-4.rq
14:47:12 <LeeF> data-diamond-loop.ttl
14:47:45 <AndyS> :p{2,}
14:47:59 <sandro> AndyS: If you look at that query, the core of it is :p{2,}
14:48:04 <LeeF> a -> b -> z      a -> c -> z       c -> c
14:48:14 <sandro> ... part of it is an unbounded operation
14:48:27 <sandro> ... they don't behave the same in how they treat cardinalities
14:48:37 <sandro> ... so we can't reverse the operation
14:48:54 <LeeF> diamond-loop-6.srx
14:48:58 <sandro> ... my implementation was doing it the wrong way around.
14:49:08 <LeeF> c / a / a / a
14:49:17 <LeeF> should be
14:49:23 <LeeF> c / c / a / a / a
14:49:35 <sandro> AndyS: Correct result is c / c / a / a / a
14:50:38 <sandro> Lee: I'm updating srx file now.
14:51:00 <LeeF> http://www.w3.org/2009/sparql/docs/tests/data-sparql11/property-path/diamond-loop-6.srx
14:52:03 <sandro> Zakim, list attendees
14:52:03 <Zakim> As of this point the attendees have been SteveH, AndyS, +1.310.729.aaaa, kasei, LeeF, MattPerry, Sandro, bglimm, pgearon, Souri_
14:52:43 <sandro> kasei: I havent looked at 34 or 35, but I have passed them.   
14:52:49 <sandro> lee: that's good enough, I think.
14:52:56 <LeeF> PROPOSED: Re-approve :pp29 with change made in this teleconference and approve :pp34 and :pp35
14:53:07 <AndyS> seconded
14:53:18 <LeeF> RESOLVED: Re-approve :pp29 with change made in this teleconference and approve :pp34 and :pp35
14:54:41 <LeeF> PROPOSED: Approve all of the tests in syntax-query, syntax-update-1, syntax-update-2
14:54:47 <kasei> +1
14:54:58 <LeeF> RESOLVED: Approve all of the tests in syntax-query, syntax-update-1, syntax-update-2
14:55:07 <LeeF> ACTION: Lee to mark property path and syntax tests as approved
14:55:07 <trackbot> Created ACTION-436 - Mark property path and syntax tests as approved [on Lee Feigenbaum - due 2011-04-12].
14:55:59 <sandro> ADJOURN
14:56:02 <bglimm> bye
14:56:03 <AndyS> 54, 52, 1 tests
14:56:04 <MattPerry> bye
14:56:04 <Zakim> -Sandro
14:56:06 <Zakim> -LeeF
14:56:06 <Zakim> -AndyS
14:56:07 <Zakim> -bglimm
14:56:07 <Zakim> -SteveH
14:56:11 <Zakim> -kasei
14:56:18 <Zakim> -pgearon
14:56:20 <Zakim> -MattPerry
14:56:53 <kasei> LeeF, are you going to update the test manifests to mark approved tests?
# SPECIAL MARKER FOR CHATSYNC.  DO NOT EDIT THIS LINE OR BELOW.  SRCLINESUSED=00000204