Chatlog 2012-05-29

See original RRSAgent log and preview nicely formatted version.

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

<kasei> Present: LeeF, kasei, AndyS, pgearon, swh, Sandro, cbuilara, MattPerry, Chimezie, Arthur, bglimm, EricP, AxelPolleres
13:58:01 <RRSAgent> RRSAgent has joined #sparql
13:58:01 <RRSAgent> logging to http://www.w3.org/2012/05/29-sparql-irc
13:58:03 <trackbot> RRSAgent, make logs world
13:58:03 <Zakim> Zakim has joined #sparql
13:58:05 <trackbot> Zakim, this will be 77277
13:58:06 <trackbot> Meeting: SPARQL Working Group Teleconference
13:58:06 <trackbot> Date: 29 May 2012
13:58:07 <Zakim> ok, trackbot; I see SW_(SPARQL)10:00AM scheduled to start in 2 minutes
13:58:53 <Zakim> SW_(SPARQL)10:00AM has now started
13:59:00 <Zakim> +LeeF
13:59:10 <Zakim> +??P1
13:59:19 <Zakim> +kasei
13:59:19 <AndyS> zakim, ??P1 is me
13:59:20 <Zakim> +AndyS; got it
14:00:39 <Zakim> +pgearon
14:00:47 <Arthur> Arthur has joined #sparql
14:01:11 <Zakim> +??P11
14:01:16 <swh> Zakim, ??P11 is me
14:01:16 <Zakim> +swh; got it
14:02:14 <cbuilara> cbuilara has joined #sparql
14:02:20 <MattPerry> MattPerry has joined #sparql
14:02:35 <Zakim> +Sandro
14:02:37 <chimezie> chimezie has joined #sparql
14:02:52 <Zakim> +??P22
14:02:56 <cbuilara> zakim, ??P22 is me
14:03:07 <Zakim> +cbuilara; got it
14:03:13 <Zakim> +MattPerry
14:03:34 <LeeF> Chair: LeeF
14:03:36 <LeeF> Regrets: Axel
14:04:06 <Zakim> +Chimezie
14:04:34 <kasei> scribenick: kasei
14:04:47 <LeeF> Regrets: Axel, Olivier
14:04:55 <LeeF> topic: Admin
14:04:57 <bglimm> bglimm has joined #sparql
14:05:01 <LeeF> PROPOSED: Accept minutes from http://www.w3.org/2009/sparql/meeting/2012-05-22
14:05:16 <Zakim> +Arthur
14:05:26 <Zakim> +??P37
14:05:32 <bglimm> Zakim, ??P37 is me
14:05:32 <Zakim> +bglimm; got it
14:05:33 <LeeF> RESOLVED: Accept minutes from http://www.w3.org/2009/sparql/meeting/2012-05-22
14:05:39 <AndyS> +1 to minutes
14:05:58 <AxelPolleres> AxelPolleres has joined #sparql
14:06:01 <kasei> LeeF: next week is during semtech
14:06:09 <AndyS> and is UK holiday
14:06:20 <kasei> ... who's going to be there?
14:06:27 <AndyS> (Tuesday - not the whole week!)
14:06:42 <kasei> ... let's plan on a short telecon.
14:07:04 <kasei> ... will check in on actions, try to make progress where we can.
14:07:46 <LeeF> topic: Document Status and To PR
14:07:34 <kasei> LeeF: greg sent email out summarizing test status.
14:07:49 <kasei> ... want to fill in detail on To_PR page. will use email as a guide.
14:07:49 <LeeF> http://www.w3.org/2009/sparql/wiki/To_PR
14:08:03 <LeeF> http://lists.w3.org/Archives/Public/public-rdf-dawg/2012AprJun/0182.html
14:08:38 <kasei> LeeF: starting with update.
14:09:08 <kasei> pgearon: not entirely caught up on update issues.
14:09:38 <kasei> LeeF: three questions: what's changed since LC? who will take responsibility for Kjetil's comment? what editorial changes are left?
14:09:48 <kasei> ... can we come back to update later in the call?
14:09:56 <kasei> pgearon: what comment from Kjetil?
14:10:02 <LeeF> http://lists.w3.org/Archives/Public/public-rdf-dawg-comments/2012May/0023.html
14:10:02 <AxelPolleres> FWIW, bnode reuse across INSERT patterns is solved. 
14:10:26 <AxelPolleres> I have drafted a reply for that one, I think.
14:10:30 <LeeF> Subtopic: Protocol
14:10:49 <AxelPolleres> http://www.w3.org/2009/sparql/wiki/CommentResponse:KK-15
14:11:05 <kasei> LeeF: haven't made any changes to Protocol since LC.
14:11:27 <kasei> ... don't need specific reviewers unless we make new changes.
14:11:42 <LeeF> http://www.w3.org/2009/sparql/wiki/Comments
14:12:05 <kasei> ... there are no open protocol comments.
14:12:09 <kasei> ... no open issues.
14:13:12 <LeeF> http://www.w3.org/2009/sparql/docs/protocol-1.1/Overview.xml
14:13:30 <AndyS> I see comments (delayed RC-1, RC-2)
14:13:38 <kasei> ... we had added text indicating UTF-8 is the only valid charset for query via POST directly.
14:13:46 <kasei> ... we don't say the same thing for update via POST directly.
14:14:00 <kasei> ... I think this was just an oversight, and we should say so for both query and update.
14:14:20 <kasei> ... does anybody have any issues with that?
14:14:31 <AndyS> ... RC-1, RC-2 MLR-1 
14:15:03 <AndyS> (old, postponed - may have been answered but not recorded?)
14:15:29 <kasei> LeeF: need to respond to old comments RC-1, RC-2, MLR-1.
14:16:15 <kasei> ... I'll draft responses to these.
14:17:18 <LeeF> ACTION: Lee to draft responses to RC-1, RC-2, and MLR-1
14:17:18 <trackbot> Created ACTION-627 - Draft responses to RC-1, RC-2, and MLR-1 [on Lee Feigenbaum - due 2012-06-05].
14:17:39 <LeeF> q?
<kasei> subsubtopic: testing
14:18:20 <AndyS> kasei: list of tests for protocol -- work-in-progress
14:18:28 <LeeF> https://gist.github.com/2770483
14:18:34 <AndyS> ... impl a validator
14:18:43 <AndyS> ... challenges around update
14:18:55 <kasei> http://kasei.us/2009/09/sparql/protocol_validator.cgi
14:19:44 <AndyS> Leef: where does the dataset come from?
14:20:15 <AndyS> kasei: end point can refuse with and also without a dataset
14:20:26 <AndyS> ... makes testing tricky
14:21:15 <AndyS> (offline to check tests against Fuseki)
14:21:31 <AndyS> LeeF: collect endpoints and deal with what we get.
14:22:08 <AndyS> kasei: OK for query ... but update is harder
14:22:22 <AndyS> LeeF: Ask for targets to be provided
14:22:59 <AndyS> ... can put up a endpoint for Anzo.
14:23:34 <AndyS> ... like query, ask for implementers to contact WG (off list!) for testing points.
14:23:51 <kasei> LeeF: will need to approve the tests.
14:24:06 <AndyS> ... need to do: complete tests, approve tests
14:24:20 <AndyS> kasei: Some tests need refining
14:24:55 <AndyS> ... currently (hesitantly) fairly complete
14:25:06 <AndyS> LeeF: I will review the tests
14:25:08 <kasei> LeeF: if you can finish specing out the tests on the github page...
14:25:43 <kasei> ... work with Andy to get a Fuseki endpoint that we can run it against, and yours. Maybe Anzo as well.
14:25:45 <AndyS> ... for approving, RDF:Query ep, Anzo ap, Jena ap
14:26:10 <AndyS> ... validate the validator
14:27:02 <LeeF> ACTION: Greg to complete specification of protocol tests at https://gist.github.com/2770483
14:27:03 <trackbot> Created ACTION-628 - Complete specification of protocol tests at https://gist.github.com/2770483 [on Gregory Williams - due 2012-06-05].
14:27:13 <LeeF> ACTION: Lee to review protocol test specifications pending completion of ACTION-628
14:27:14 <trackbot> Created ACTION-629 - Review protocol test specifications pending completion of ACTION-628 [on Lee Feigenbaum - due 2012-06-05].
14:27:29 <LeeF> ACTION: Greg to complete implementation of validator pending completion of ACTION-628
14:27:29 <trackbot> Created ACTION-630 - Complete implementation of validator pending completion of ACTION-628 [on Gregory Williams - due 2012-06-05].
14:27:58 <LeeF> ACTION: Greg to work with Andy (and possibly Lee, others) to check validator with existing protocol implementations pending completion of ACTION-630
14:27:58 <trackbot> Created ACTION-631 - Work with Andy (and possibly Lee, others) to check validator with existing protocol implementations pending completion of ACTION-630 [on Gregory Williams - due 2012-06-05].
14:28:15 <kasei> LeeF: once we've done these actions, should be able to approve protocol tests.
14:28:16 <LeeF> ACTION-631?
14:28:16 <trackbot> ACTION-631 -- Gregory Williams to work with Andy (and possibly Lee, others) to check validator with existing protocol implementations pending completion of ACTION-630 -- due 2012-06-05 -- OPEN
14:28:16 <trackbot> http://www.w3.org/2009/sparql/track/actions/631
14:29:11 <kasei> ... does Fuseki implement all the ways of sending queries?
14:29:14 <kasei> AndyS: yes.
14:29:54 <kasei> kasei: I support all methods as well.
14:30:01 <kasei> LeeF: I'm not sure we support them all in Anzo yet.
14:30:12 <kasei> ... anybody else?
14:30:28 <swh> pretty sure ours is not complete
14:30:30 <swh> either of them
<kasei> subtopic: Service Description
14:30:52 <LeeF> kasei: no significant changes since last call
14:31:24 <AndyS> kasei: no editorial known.
14:31:34 <AndyS> ... one comment (Chime drafted)
14:32:32 <kasei> LeeF: no ack on MO-1?
14:32:34 <kasei> kasei: no.
14:33:45 <kasei> LeeF: I'll send a ping on MO-1.
14:34:37 <kasei> ... sent.
14:35:22 <kasei> ... no ack from TI-3.
14:35:29 <kasei> ... will send a ping on TI-3, also.
14:36:53 <kasei> ... KR-1?
14:37:12 <kasei> ... will follow up offlist to clarify status.
14:37:48 <kasei> ... all comments have been responded to.
14:38:06 <kasei> ... no issues.
14:38:28 <LeeF> kasei: i have the SD validator
14:38:30 <LeeF> ... conformance is minimal
14:38:35 <LeeF> ... we only have my implementation right now
14:38:42 <LeeF> ... there might be others, and conformant support is trivial
14:38:54 <kasei> LeeF: any other implementations of SD?
14:39:10 <kasei> ... what about validating the validator? approving the tests?
14:39:16 <AndyS> Validator is where?
14:39:17 <chimezie> I have a bare bones implementation of SD
14:39:46 <kasei> http://kasei.us/2009/09/sparql/sdvalidator.cgi
14:39:59 <kasei> https://github.com/kasei/sparql11-sdvalidator
14:40:20 <kasei> chimezie: I have a basic SD implementation.
14:40:51 <LeeF> ACTION: Greg to document what the service description validator tests
14:40:51 <trackbot> Created ACTION-632 - Document what the service description validator tests [on Gregory Williams - due 2012-06-05].
14:40:53 <kasei> LeeF: is there any documentation for the SD validator?
14:40:57 <kasei> kasei: no. I'll add some.
14:41:12 <LeeF> ACTION: Chimezie to put his SD implementation public somewhere & try the SD validator against it
14:41:13 <trackbot> Created ACTION-633 - Put his SD implementation public somewhere & try the SD validator against it [on Chimezie Ogbuji - due 2012-06-05].
14:41:55 <LeeF> ACTION: Lee to look over the SD validator code (https://github.com/kasei/sparql11-sdvalidator/blob/master/sdvalidator.cgi)
14:41:55 <trackbot> Created ACTION-634 - Look over the SD validator code (https://github.com/kasei/sparql11-sdvalidator/blob/master/sdvalidator.cgi) [on Lee Feigenbaum - due 2012-06-05].
14:42:02 <LeeF> ACTION-632
14:42:04 <LeeF> ACTION-632?
14:42:04 <trackbot> ACTION-632 -- Gregory Williams to document what the service description validator tests -- due 2012-06-05 -- OPEN
14:42:04 <trackbot> http://www.w3.org/2009/sparql/track/actions/632
14:43:03 <kasei> LeeF: does anybody else have plans to implement SD in coming weeks?
14:43:11 <swh> I have an impl. but it's out of date
14:43:16 <swh> might be able to update it
14:43:27 <AndyS> Hope to tweak things but no guarantees
14:43:44 <kasei> LeeF: suspect ericP might have an implementation of SD.
14:43:56 <kasei> sandro: I don't think so. Possible, but haven't heard him mention it.
14:44:01 <kasei> LeeF: I'll ping him about it.
14:44:16 <Arthur> We plan to implement, but are a few months away
14:44:43 <LeeF> thanks, Arthur
<kasei> subtopic: Graph Store Protocol
14:44:45 <kasei> LeeF: regarding GSP, chime drafted responses to outstanding comments.
14:45:01 <kasei> ... can somebody take a look at chime's drafts?
14:45:20 <kasei> kasei: I did this morning. Sent a comment about it.
14:47:20 <ericP> huh, i don't think i have much SD outside of a little demo mode which claims something to be a DOAP project
14:47:26 <ericP> need one for CR?
14:47:28 <LeeF> ok, thanks ericP
14:47:29 <kasei> LeeF: maybe we remove the word "REST"?
14:47:36 <LeeF> ericP, maybe but probably not
14:47:49 <kasei> chimezie: the more important part of the response is that what's being asked for is out of scope for the GSP.
14:47:50 <Zakim> +EricP
14:48:13 <kasei> chimezie: I'm not sure why it's not RESTful.
14:50:11 <kasei> kasei: this violates HATEOAS
14:50:35 <ericP> chimezie, as soon as you define a API then you've violated the generalized mechanism of HTTP
14:51:12 <kasei> chimezie: we've tied our own hands on this, because we've decided service description is out of scope for GSP.
14:51:43 <kasei> LeeF: we're not doing full REST because it's not practical at this point. we do have an HTTP API.
14:52:00 <kasei> ... do we remove any mention of REST? or do we add a note saying it's "REST-ish", but not actually REST?
14:52:28 <kasei> sandro: the SD gives you a URI for the dataset.
14:52:52 <kasei> ... why don't we take out indirect reference?
14:52:55 <AndyS> q+ to look at where REST is used in doc
14:53:10 <kasei> ... in practice, you'd have to not use graph names outside of your control (which is a best practice anyway)
14:53:27 <kasei> chimezie: I think parts of REST are dogma.
14:53:47 <LeeF> q+ to respond to Sandro's suggestion
14:53:57 <kasei> ... the question of HATEOAS is dogmatic.
14:54:11 <kasei> sandro: i've beent hinking of what are good ways to deal with datasets.
14:54:26 <kasei> ... if you're using somebody else's URI as a graph name, if somebody wants to find out what's there, they might as well ask the someone else.
14:54:32 <AndyS> I see it used in   abstract, intro, terminology, PATCH, references
14:54:38 <kasei> ... if they don't want to do REST, they should use SPARQL.
14:54:45 <swh> This is a usecase we have now
14:55:57 <kasei> sandro: the dataset uri is something they pick.
14:56:13 <kasei> AndyS: can we discuss where "REST" is used in the document?
14:56:40 <kasei> LeeF: I think a lot of people want to be able to use HTTP operations on a graph store which has graphs which may not have http URIs or be on the same host as the GSP implementation.
14:57:01 <swh> +1 to LeeF
14:57:19 <kasei> sandro: you wouldn't have this architectural problem if you used SPARQL instead of GSP for those use cases.
14:57:30 <kasei> ... practically it would be much easier, because you wouldn't have to get the SD.
14:57:56 <kasei> LeeF: because of time, let's not make a decision on this now.
14:58:08 <kasei> ... we'll try to take this up again next week or the week after.
14:58:36 <kasei> AndyS: only found REST in abstract, intro, a few other places. Not used very much.
14:58:40 <sandro> NO --  I'm not saying use SPARQL , I'm saying make indirect reference parse of SPARQL Protocol.
14:58:42 <kasei> LeeF: who would object to removing the term "REST"?
14:59:32 <kasei> chimezie: the term is used minimally to avoid this type of angels on a pinhead conversation.
14:59:50 <kasei> ... if you don't refer to REST it makes it harder for a developer to comprehend.
15:00:26 <kasei> AndyS: I have sympathy for that argument. "hardcore REST" vs. "REST style".
15:00:40 <kasei> ... we could appeal to HTTP rather than saying it's part of "REST".
15:00:55 <kasei> ... we want "do the natural thing in accordance with good style"
15:00:55 <sandro> +1 andy -- just reference HTTP not REST, as a way through this, maybe....
15:00:56 <swh> +1 to AndyS 
15:01:17 <kasei> ... don't want to get into arguments about having to use http URIs for things, because there are reasons to use urns, etc.
15:01:33 <kasei> LeeF: Andy's suggestion is very appealing to me.
15:02:03 <kasei> ... share chimezie's worldview, but it may be the best compromise to appeal to HTTP instead of REST.
15:02:22 <kasei> AndyS: long-running situation which led to HATEOAS.
15:02:53 <kasei> chimezie: it makes sense. the parts of REST I think of as engineering principles correspond one-to-one with HTTP.
15:03:28 <kasei> ... I'm convinced that it's the prudent thing to do. Emphasize conformance to HTTP, remove "REST".
15:03:46 <kasei> sandro: should we mention "REST" as "some people might call this REST"?
15:04:03 <bglimm> sorry, I have another meeting now and can't stay longer... talk to you next week
15:04:23 <kasei> chimezie: unless not mentioning REST would cause somebody to object, I don't see the value in keeping it in.
15:04:35 <kasei> LeeF: upside is that if somebody searches for "SPARQL REST", they're more likely to find the document.
15:04:46 <kasei> sandro: but somebody will write a blog post that they find anyway.
15:05:06 <kasei> LeeF: mentioning REST will cause people to have the concerns they've already expressed.
15:05:17 <kasei> chimezie: would that be an editorial change?
15:05:23 <kasei> LeeF: I think the WG would consider that editorial.
15:05:53 <kasei> ... sandro, would you like to still explore the indirect identification stuff?
15:05:55 <kasei> sandro: no.
15:06:10 <kasei> LeeF: if we had a resolution about this, we should have a new one.
15:06:21 <kasei> ... but that shouldn't stop you from rewording in the document.
15:06:40 <swh> sorry, have to go, late for another call
15:06:46 <LeeF> ACTION: Chimezie to draft text for GSP that replaces REST references with HTTP references and send summary/diff to the mailing list
15:06:46 <trackbot> Created ACTION-635 - Draft text for GSP that replaces REST references with HTTP references and send summary/diff to the mailing list [on Chimezie Ogbuji - due 2012-06-05].
# SPECIAL MARKER FOR CHATSYNC.  DO NOT EDIT THIS LINE OR BELOW.  SRCLINESUSED=00000263