Chatlog 2009-06-30

From SPARQL Working Group
Revision as of 08:14, 1 July 2009 by Aseaborne (Talk | contribs)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

See original RRSAgent log and preview nicely formatted version.

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

13:55:44 <trackbot> Zakim, this will be 77277
13:55:44 <Zakim> ok, trackbot; I see SW_(SPARQL)10:00AM scheduled to start in 5 minutes
13:55:45 <trackbot> Meeting: SPARQL Working Group Teleconference
13:55:45 <trackbot> Date: 30 June 2009
13:55:49 <LeeF> Zakim, this will be SPARQL
13:55:49 <Zakim> ok, LeeF; I see SW_(SPARQL)10:00AM scheduled to start in 5 minutes
13:55:58 <LeeF> Chair: LeeF
13:56:03 <LeeF> Scribenick: KjetilK
13:56:12 <LeeF> Agenda: http://www.w3.org/2009/sparql/wiki/Agenda-2009-06-30
13:56:12 <AndyS> Present: LeeF, AndyS, kasei, AlexPassant, bglimm, pgearon, Prateek, Orri, EricP, AxelPolleres, swh, LukeWM, KjetilK, SimonS, iv_an_ru
13:57:45 <iv_an_ru> Orri got problem with his connection so he is missing today.
13:57:47 <Zakim> SW_(SPARQL)10:00AM has now started
13:58:19 <AndyS> Request for agenda addition: under "update" to discuss whether to start with the submission as the strawman for the update language.  Are there alternatives to consider? (After strawman -- start throwing rocks at it to test it out)
13:58:35 <LeeF> AndyS, ack
14:00:21 <LeeF> Regrets: Chimezie, SimonKJ, Bijan
14:00:57 <AndyS> bglimm, if there is an example, I'd appreciate seeing it to be informed about the issues but theer is no time-critical issue here for me.
14:01:22 <bglimm> I can check the OWL tests for one, but I haven't done that yet
14:01:51 <LeeF> zakim, who's on the phone?
14:01:51 <Zakim> On the phone I see LeeF, AndyS, kasei (muted), AlexPassant, bglimm (muted)
14:05:28 <LeeF> zakim, who's on the phone?
14:05:38 <Zakim> On the phone I see LeeF, AndyS, kasei (muted), AlexPassant, bglimm (muted), pgearon, Prateek
14:06:29 <ericP> Zakim, please dial ericP-office
14:06:34 <AndyS> Scribenick: AndyS
14:06:53 <AndyS> Scribe: AndyS
14:06:56 <LeeF> PROPOSED: Approve minutes at http://www.w3.org/2009/sparql/meeting/2009-06-23
14:07:19 <bglimm> +1
14:07:54 <AndyS> RESOLVED Accepted the minutes of the meeting of 2009-06-23
14:08:35 <AxelPolleres> regrets for next week
14:08:49 <AndyS> Next meeting: July 7
14:10:05 <LeeF> Please let me know if you'll be on holiday anytime over the rest of the summer
14:10:16 <AndyS> RDB2RDB WG: Now has 2 co-chairs.
14:10:33 <AndyS> ... and will start soon.
14:11:07 <AndyS> Nothing else from liaisons
14:11:22 <AndyS> Topic: Open Actions
14:11:35 <AndyS> LeeF: Please close your open actions.
14:11:45 <AndyS> Topic: FPWD of F&R
14:12:09 <AndyS> Input for charter for phase II of SPARQL WG
14:12:32 <LeeF> PROPOSED: Publish http://www.w3.org/2009/sparql/docs/features/ version 1.114 as First Public Working Draft 
14:12:35 <AlexPassant> wrt open actions, can LeeF or AxelPolleres close http://www.w3.org/2009/sparql/track/actions/43 ? (already discussed previously)
14:13:22 <AndyS> incorporates changes from last week: text for time permitting features, language fixes, OWL Direct semantics included
14:13:50 <AxelPolleres> alex, done
14:13:51 <AndyS> Any reservations?
14:14:03 <AndyS> (none)
14:14:15 <LeeF> PROPOSED: Publish http://www.w3.org/2009/sparql/docs/features/ version 1.114 as First Public Working Draft
14:14:29 <LeeF> RESOLVED: Publish http://www.w3.org/2009/sparql/docs/features/ version 1.114 as First Public Working Draft 
14:15:01 <AndyS> No objections or abstentions
14:15:08 <AndyS> zakim, who is on the phone?
14:15:08 <Zakim> On the phone I see LeeF, AndyS, kasei (muted), AlexPassant, bglimm (muted), pgearon, Prateek, EricP, Orri, AxelPolleres, [Garlik]
14:15:10 <Zakim> [Garlik] has swh, LukeWM
14:15:17 <LeeF> ACTION: Eric to work with Lee to publish FPWD of features & rationales
14:15:17 <trackbot> Created ACTION-50 - Work with Lee to publish FPWD of features & rationales [on Eric Prud'hommeaux - due 2009-07-07].
14:15:44 <AndyS> LeeF: Editors will continue to work on it.
14:15:50 <ericP> F&R is valid HTML according to http://validator.w3.org/check?uri=http://www.w3.org/2009/sparql/docs/features/
14:16:32 <AndyS> AndyS: Comments will arrive now.
14:16:40 <AxelPolleres> q+
14:16:55 <AndyS> LeeF: We do not need to formally track at the level of CR etc.
14:16:57 <LeeF> ack AxelPolleres
14:18:27 <AndyS> Axel: Comments timeframe?
14:18:29 <AndyS> q+
14:18:45 <kasei> the F&R TOC has a 5.5, but it doesn't actually exist (?)
14:18:56 <LeeF> invisible ink?
14:18:59 <AndyS> ack me
14:19:18 <AndyS> EricP: ... a deadline to hint at timeliness
14:19:19 <kasei> 5.3 and 5.4 in invisible ink, certainly, but 5.5 shows up in the TOC :)
14:20:39 <ericP> http://validator.w3.org/checklink?url=http://www.w3.org/2009/sparql/docs/features/#results1 found a broken link to http://www.w3.org/2009/sparql/docs/features/owl.css and three broken fragments: Entailment_status, Subqueries_issues, Basic_federated_querys_status
14:21:00 <AndyS> LeeF: end August?  LC is autumn.
14:21:03 <AlexPassant> kasei: remove it from the TOC, is there a way to change the publishing action from 1.114 to 1.115 ?
14:21:31 <AxelPolleres> q+ to say why 31 august fine with me.
14:21:40 <ericP> KjetilK, AlexPassant, do you know where i should find owl.css?
14:21:57 <AlexPassant> ericP: I removed that link - was useless (comes from the OWL template we re-used)
14:22:05 <AndyS> The Scribe: Scribe notes the resolution has the wrong version now.
14:23:00 <AndyS> LeeF: Can publish before comment period because it's a WD.
14:23:23 <AndyS> EricP: Just need to note that comments may be relative to first version.
14:23:26 <LeeF> ack AxelPolleres
14:23:26 <Zakim> AxelPolleres, you wanted to say why 31 august fine with me.
14:23:52 <AndyS> Axel: Vaction season is fast approaching
14:24:06 <AndyS> q+
14:24:10 <kasei> AlexPassant: 2.8.5 needs to be dropped, too, I think.
14:24:38 <AndyS> ack me
14:25:00 <KjetilK> +1 to that
14:25:14 <AndyS> LeeF: Suggest saying comments to mailing list, no date.
14:25:18 <ericP> AlexPassant, KjetilK, where should #Subqueries_issues point?
14:25:43 <AndyS> ... new resolution ...
14:25:56 <kasei> yes
14:26:08 <kasei> (what ericP said)
14:26:34 <AndyS> (formatting/editorial discussion)
14:27:00 <KjetilK> ericP, #Subqueries_discussion probably
14:27:56 <AlexPassant> q+ to ask about last fix
14:31:06 <kasei> sure, but i have nothing to say new from two weeks ago.
14:31:08 <AxelPolleres> q?
14:31:21 <AlexPassant> done from my side as well
14:31:33 <AlexPassant> 1.118
14:31:46 <LeeF> PROPOSED: Publish http://www.w3.org/2009/sparql/docs/features/ version 1.118 as FPWD, with comments to be solicited indefinitely to public-rdf-dawg-comments@w3.org 
14:31:52 <ericP> second
14:32:00 <AxelPolleres> +1
14:32:11 <bglimm> +1
14:32:14 <LeeF> RESOLVED: Publish http://www.w3.org/2009/sparql/docs/features/ version 1.118 as FPWD, with comments to be solicited indefinitely to public-rdf-dawg-comments@w3.org 
14:32:43 <AndyS> Topic: Service descriptions
14:33:10 <AndyS> No disucssion of HTTP OPTIONs last discussion.  Kasei presented his ideas.
14:33:22 <AndyS> LeeF: Short summary?
14:33:49 <AndyS> Kasei: coin URIs.  One RDFS class for endpoints.  Properties for endpoint to details.
14:33:51 <AxelPolleres> (no objections or abstentions on the proposal to publish ... just for the records)
14:33:58 <AndyS> .. inc dataset description.
14:34:14 <AndyS> ... work with active void2 developement 
14:34:23 <LeeF> http://www.w3.org/2009/sparql/wiki/Feature:ServiceDescriptions#Strawman_Proposal_for_Service_Description_Vocabulary_and_URIs
14:34:47 <AndyS> LeeF: Who is responsible for HTTP OPTIONS proposal?
14:35:32 <AndyS> ... Jacek
14:35:42 <AxelPolleres> that was jacek, but he told me he is only attending occasionally.
14:35:49 <AndyS> ... feedback?
14:36:05 <swh> +1 to not over specifying
14:36:28 <AndyS> AndyS: Looks like a good start.
14:36:46 <SimonS> +1 Looks like the sweet spots have been hit.
14:36:48 <LeeF> zakim, who's here?
14:36:48 <Zakim> On the phone I see AndyS, kasei, AlexPassant, bglimm (muted), pgearon, Prateek, EricP, Orri, AxelPolleres, [Garlik], SimonS, LeeF, KjetilK (muted)
14:36:50 <Zakim> [Garlik] has swh, LukeWM
14:36:50 <AndyS> Steve: Going beyond is getting out of WG scope; less is not useful.
14:36:51 <Zakim> On IRC I see SimonS, LukeWM, AxelPolleres, Prateek, AndyS_, pgearon, bglimm, Zakim, RRSAgent, LeeF, kasei, AlexPassant, AndyS, swh, karl, john-l, iv_an_ru, KjetilK, trackbot, ericP
14:37:59 <bglimm> no strong feelings
14:37:59 <AndyS> AndyS: Would like to proced with this.
14:38:23 <AndyS> q+ to ask about finding descriptions
14:38:24 <AxelPolleres> q+ to ask about liaison with "void"
14:38:31 <AlexPassant> q-
14:39:08 <KjetilK> ack AndyS
14:39:08 <Zakim> AndyS, you wanted to ask about finding descriptions
14:40:03 <kasei> agree with AndyS
14:41:04 <AxelPolleres> 2 big blocks: (i) service descriptions (what?) (ii) how to find/serve them (how?)
14:41:24 <AndyS> AndyS: thoughts on how to find descriptions.  3rd party matters a lot but what about services publishing info about themselves?
14:41:59 <kasei> does anyone have thoughts on the "extra" features I list as possibilities in the strawman, especially the first two?
14:42:01 <AndyS> LeeF: Strawman kasei's proposal and start on refinement/discussion.
14:42:05 <swh> q+ to advocate HTTP headers
14:42:17 <AxelPolleres> ack me
14:42:17 <Zakim> AxelPolleres, you wanted to ask about liaison with "void"
14:42:39 <AndyS> Axel: Liaison with void?  Formally?
14:43:13 <SimonS> +q to comment on kaseis extra features
14:43:13 <kasei> Michael Hausenblas asked me to be the point of contact with them, but as mentioned it's less than formal.
14:43:15 <swh> I disagree that we should liase
14:43:41 <AndyS> Be good for someone to keep the WG up-to-date.
14:44:00 <ericP> void is in a space which is likely to shift a whole lot. we should consider our job to link to things playing the void role
14:44:29 <AndyS> swh: Don't sanction one schema.  Instead, be neutral - don't refer to outside vocabs.
14:44:37 <ericP> i don't mind using them in examples
14:44:44 <swh> yes, examples is ok
14:45:05 <LeeF> PROPOSED: Adopt kasei's strawman proposal of what goes into a service description (http://www.w3.org/2009/sparql/wiki/Feature:ServiceDescriptions#Strawman_Proposal_for_Service_Description_Vocabulary_and_URIs) and begin to fill in details
14:45:30 <ericP> seocnd
14:45:30 <AxelPolleres> +1
14:45:33 <SimonS> second
14:45:35 <KjetilK> +1
14:45:44 <bglimm> +1
14:45:44 <LeeF> RESOLVED: Adopt kasei's strawman proposal of what goes into a service description (http://www.w3.org/2009/sparql/wiki/Feature:ServiceDescriptions#Strawman_Proposal_for_Service_Description_Vocabulary_and_URIs) and begin to fill in details
14:45:46 <AndyS> No abstentions or objections
14:46:07 <SimonS> q-
14:46:33 <LeeF> ACTION: Greg to flesh out service description proposal with initial URI suggestions for features, classes, properties, etc.
14:46:34 <trackbot> Created ACTION-51 - Flesh out service description proposal with initial URI suggestions for features, classes, properties, etc. [on Gregory Williams - due 2009-07-07].
14:47:01 <swh> ack me
14:47:01 <Zakim> swh, you wanted to advocate HTTP headers
14:47:01 <AndyS> Topic: Finding service descriptions
14:47:29 <AndyS> swh: Use HTTP headers pointing to a URI.
14:47:52 <AndyS> ... but relative URI needed because sevice does not know what it's external name is
14:48:15 <kasei> it's in the wiki under existing impls., not the strawman.
14:48:42 <LeeF> option 1 - use HTTP header in any GET/HEAD request on the endpoint (with or without a query)
14:48:55 <LeeF> option 2 - use HTTP OPTIONS verb
14:48:59 <AxelPolleres> Steve says ANY http get reauest should have a pointer to the description?
14:49:00 <AndyS> ... URI is the endpoint with or without ?query
14:49:11 <LeeF> option 3 - place service description RDF in a well-known location on the server
14:49:19 <LeeF> option 4 - DESCRIBE query against SPARQL endpoint URI
14:49:22 <ericP> q+ to suggest that [[ SELECT ?foo ?bar { <srvc> <x> ?foo ; <y> ?bar } ]] requires no extension to HTTP tools like apache, iis, ...
14:49:39 <pgearon> +q
14:49:54 <LeeF> option 5 - suggest that endpoints SHOULD answer queries about their own URI
14:49:57 <AndyS> q+ to talk about endpoint URIs and separation from query engine
14:50:20 <swh> what if they don't know their own URI? or the client doesn't know the real one
14:50:26 <LukeWM> q+
14:50:27 <AndyS> ack EricP
14:50:27 <Zakim> ericP, you wanted to suggest that [[ SELECT ?foo ?bar { <srvc> <x> ?foo ; <y> ?bar } ]] requires no extension to HTTP tools like apache, iis, ...
14:51:00 <AndyS> EricP: Make dft graph have own description in it.
14:51:14 <AndyS> pgearon: Not so sure about thisl
14:51:38 <ericP> but the SPARQL query *is* a GET
14:51:39 <AndyS> ... not RESTful.  GET on a web resource.
14:51:51 <swh> q+ to say you can use 4 via 1
14:52:24 <AxelPolleres> ack pgearon
14:52:29 <kasei> another option, which I think no one is going to care for, is a new sparql query method
14:52:30 <AndyS> ack AndyS
14:52:30 <Zakim> AndyS, you wanted to talk about endpoint URIs and separation from query engine
14:52:34 <ericP> HEAD is the same as GET without the response body
14:53:07 <swh> +1 to AndyS on seperation
14:53:32 <pgearon> ericP, yes, in general, but when used in REST then it is usually used to retrieve metadata
14:53:52 <pgearon> HTTP only says that HEAD *should* start the same as GET
14:54:23 <AndyS> AndyS: issue can be that the protocol and dataset are separate in the service impl
14:56:01 <LeeF> q?
14:56:25 <AndyS> ack LukeWM
14:56:38 <AndyS> LukeWM: option 1 pref
14:56:47 <AndyS> ... avoids multipel requests
14:56:50 <kasei> q+ to worry about difficulty in using OPTIONS
14:57:20 <AndyS> ... creatively, if a service changes frequently, may be an overhead
14:57:21 <ericP> i don't see that there are more reqs required if you query the service than if you HEAD the service
14:57:28 <LeeF> ack swh
14:57:28 <Zakim> swh, you wanted to say you can use 4 via 1
14:57:44 <AndyS> swh: Can do opt 4 with canned describe query.
14:58:00 <ericP> in fact, the contrary, as querying the service gives you the data, where HEADing the service is an extra indirect
14:58:04 <AndyS> ... Garlik store would find it hard to do opt 4
14:58:20 <kasei> Zakim, unmute me
14:58:20 <Zakim> kasei should no longer be muted
14:58:39 <AndyS> ... differentiates between opt 1 and 4.
14:59:00 <kasei> having trouble keeping track of which option belongs to which number :\
14:59:19 <LeeF> i'll repeat :)
14:59:24 <AndyS> ... server has broadbased info
14:59:56 <AndyS> (Joseki does not have access to dynamic extensions the query engine may provide)
15:00:07 <LeeF> option 1 - use HTTP header in any GET/HEAD request on the endpoint (with or without a query)
15:00:07 <LeeF> option 2 - use HTTP OPTIONS verb
15:00:07 <LeeF> option 3 - place service description RDF in a well-known location on the server
15:00:07 <LeeF> option 4 - DESCRIBE query against SPARQL endpoint URI
15:00:07 <LeeF> option 5 - suggest that endpoints SHOULD answer queries about their own URI
15:00:23 <LukeWM> ericP, I think it depends on how your application works.  If you want to make the query and find out the service description whatever the description, using GET results in fewer requests.
15:00:36 <AndyS> kasei: OPTIONs has costs / difficulties.  client libraries can make it tricky.
15:00:41 <swh> +1 to OPTIONS being difficult
15:00:43 <kasei> q+
15:00:55 <AndyS> LeeF: will send options to the email list
15:01:02 <AxelPolleres> +1  with concern on OPTIONS option
15:01:26 <LeeF> q?
15:01:30 <LeeF> ack kasei
15:01:30 <Zakim> kasei, you wanted to worry about difficulty in using OPTIONS and to 
15:01:42 <AndyS> LeeF: Next week: Negation, update from strawmans on negation update submission.
15:01:44 <LeeF> option 6 - add a new SPARQL query method alongside SELECT/CONSTRUCT/ASK/DESCRIBE
15:02:03 <AxelPolleres> isn't that just DESCRIBE <endpoint> ?
15:02:04 <KjetilK> DESCRIBE ENDPOINT
15:02:04 <ericP> SELFDESCRIBE ?
15:02:15 <pgearon> I'm fine with option 6, but would still want a REST approach that was equivalent
15:02:23 <AndyS> LeeF: Thx to editors for FPWD.
15:02:25 <kasei> pgearon++
15:02:26 <Zakim> -Orri
15:02:29 <Zakim> -[Garlik]
15:02:30 <Zakim> -KjetilK
15:02:30 <Zakim> -AxelPolleres
15:02:32 <AndyS> Thx to Eric in advance
15:02:32 <Zakim> -EricP
15:02:32 <Zakim> -bglimm
15:02:33 <Zakim> -LeeF
15:02:33 <Zakim> -AlexPassant
15:02:35 <Zakim> -kasei
15:02:38 <Zakim> -Prateek
15:02:40 <Zakim> -pgearon
15:02:42 <Zakim> -SimonS
15:02:48 <AndyS> ADJOURNED
15:03:30 <kasei> I think it has a lot of issues :)
# SPECIAL MARKER FOR CHATSYNC.  DO NOT EDIT THIS LINE OR BELOW.  SRCLINESUSED=00000353