Chatlog 2010-08-10

From SPARQL Working Group
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.

13:59:57 <trackbot> Meeting: SPARQL Working Group Teleconference
13:59:57 <trackbot> Date: 10 August 2010
13:59:57 <Zakim> ok, trackbot; I see SW_(SPARQL)10:00AM scheduled to start in 1 minute
14:00:05 <LeeF> zakim, this will be SPARQL
14:00:05 <Zakim> ok, LeeF, I see SW_(SPARQL)10:00AM already started
14:00:07 <LeeF> Chair: LeeF
14:00:18 <LeeF> Agenda:
14:00:22 <Zakim> +[IPcaller]
14:00:40 <AndyS> zakim, [IPCaller] is me
14:00:40 <Zakim> +AndyS; got it
14:00:40 <NicholasH> zakim: who is on the call?
14:00:45 <SteveH_> bglimm, I got the last line I guess, sorry!
14:00:46 <Zakim> +pgearon
14:00:47 <Zakim> + +1.617.245.aaaa
14:00:47 <Zakim> +Sandro
14:00:53 <ivan> zakim, dial ivan-voip
14:00:53 <Zakim> ok, ivan; the call is being made
14:00:55 <Zakim> +Ivan
14:01:11 <LeeF> zakim, who's on the phone?
14:01:11 <Zakim> On the phone I see ??P0, Garlik, AndyS, pgearon, +1.617.245.aaaa, Sandro, Ivan
14:01:20 <LeeF> zakim, aaaa is me
14:01:20 <Zakim> +LeeF; got it
14:01:24 <SteveH> Zakim, Garlik is temporarily me
14:01:24 <Zakim> +SteveH; got it
14:01:31 <NicholasH> zakim, who is here?
14:01:31 <Zakim> On the phone I see ??P0, SteveH, AndyS, pgearon, LeeF, Sandro, Ivan
14:01:32 <Zakim> On IRC I see RRSAgent, SteveH, AxelPolleres, NicholasH, bglimm, LeeF, AndyS, ivan, karl, iv_an_ru, pgearon, sandro, trackbot, kasei
14:01:42 <sandro> sandro has changed the topic to: Agenda:
14:02:07 <NicholasH> Zakim, ??P0 is me
14:02:07 <Zakim> +NicholasH; got it
14:02:13 <ivan> zakim, mute me
14:02:23 <Zakim> Ivan should now be muted
14:02:27 <ivan> just a minute
14:02:31 <Zakim> +bglimm
14:02:42 <LeeF> topic: admin
14:02:50 <bglimm> Zakim, mute me
14:02:54 <LeeF> PROPOSED: Approve minutes at
14:02:59 <Zakim> bglimm should now be muted
14:03:20 <ivan> zakim, unmute me
14:03:20 <Zakim> Ivan should no longer be muted
14:03:40 <Zakim> +AxelPolleres
14:04:05 <AndyS> 1/ not typing, 2/ muted
14:04:12 <ivan> scribenick: ivan
14:04:37 <ivan> lee: propose to approve the minutes for last week
14:04:42 <ivan> ... carried once
14:04:47 <ivan> .... carried twice
14:04:48 <LeeF> RESOLVED: Approve minutes at
14:04:49 <AxelPolleres> seconded
14:04:51 <pgearon> +1
14:04:58 <ivan> Topic: next meeting
14:05:00 <LeeF> Next regular meeting: 2010-08-17 @ 15:00 UK / 10:00 EDT
14:05:05 <ivan> next week, same time and place
14:05:10 <AndyS> My regrets for next week's meeting.
14:05:22 <ivan> ivan: I am not 100% sure to be around next week
14:05:23 <chimezie> chimezie has joined #sparql
14:05:27 <AxelPolleres> think Alex sent regrets (on vacation)
14:05:34 <LeeF> Regrets: Nico, Greg, Alex, Chime
14:05:36 <chimezie> Zakim, what is the passcode?
14:05:36 <Zakim> the conference code is 77277 (tel:+1.617.761.6200 tel:+ tel:+44.203.318.0479), chimezie
14:06:10 <ivan> Topic: go through open issues
14:06:18 <Zakim> + +1.216.636.aabb
14:06:22 <ivan> Lee: there are some issues we may want to close, see the agenda
14:06:31 <chimezie> Zakim, +1.216.636.aabb is me
14:06:31 <Zakim> +chimezie; got it
14:06:40 <AxelPolleres> chime, sandro, could either of you look into and draft a reply?
14:07:05 <ivan> lee: any other topic for the agenda
14:07:12 <sandro> AxelPolleres, I don't think that should be me, no.
14:07:12 <chimezie> yeah, i was looking at that
14:07:18 <ivan> AxelPolleres: an additional comment from timbl, we may want to reply
14:07:36 <ivan> lee: it is better if chime looks at it, and see if there is a wg attention he may bring it up on the list
14:08:11 <AxelPolleres> I'll put chime responsible on the comments page, thanks!
14:08:16 <LeeF> topic: Issue Rodeo
14:08:36 <LeeF>
14:08:42 <chimezie> Zakim, mute me
14:08:42 <Zakim> chimezie should now be muted
14:08:48 <ivan> start with issue 48
14:08:54 <LeeF> ISSUE-48?
14:08:54 <trackbot> ISSUE-48 -- Is DELETE too verbose? -- open
14:08:54 <trackbot>
14:09:04 <sandro> no description.  :-/
14:09:25 <ivan> lee: issue 48 not repeating the pattern when the only thing you want is to delete a pattern
14:09:35 <LeeF> DELETE WHERE { ... }
14:09:47 <ivan> ... the decision a few month ago was to have delete where {...}
14:09:58 <ivan> ... we already discussed it and resolved it
14:10:02 <LeeF> PROPOSED: Resolve ISSUE-48 via current DELETE WHERE { tempalte } abbreviation as in
14:10:03 <ivan> ... the proposal is to close the issue
14:10:15 <AxelPolleres> +1
14:10:24 <SteveH> +1
14:10:24 <ivan> s/tempalte/template/
14:10:27 <ivan> +1
14:10:31 <pgearon> +1
14:10:32 <chimezie> +1
14:10:47 <NicholasH> +1
14:10:52 <sandro> +1
14:10:56 <AndyS> +1
14:10:56 <bglimm> +1
14:10:57 <LeeF> RESOLVED: Resolve ISSUE-48 via current DELETE WHERE { tempalte } abbreviation as in
14:11:06 <LeeF> trackbot, close ISSUE-48
14:11:06 <trackbot> ISSUE-48 Is DELETE too verbose? closed
14:11:11 <ivan> s/tempalte/template/
14:11:13 <LeeF> ISSUE-49?
14:11:13 <trackbot> ISSUE-49 -- Is a graph an information resource -- open
14:11:13 <trackbot>
14:11:24 <chimezie> Zakim, unmute me
14:11:24 <Zakim> chimezie should no longer be muted
14:11:33 <ivan> lee: this is something kjetill raised a while ago, chime added section 8 in the draft
14:11:40 <sandro> link to section 8 ?
14:11:42 <ivan> ... it explaines http-range
14:11:46 <LeeF> PROPOSED: Close ISSUE-49 based on the current text in Section 8 at
14:11:48 <ivan> s/explaines/explains
14:12:09 <ivan> lee: chime, is that correct?
14:12:14 <ivan> chimezie: yes, it does
14:12:32 <ivan> sandro: I have not had the time to read it
14:12:51 <ivan> ... is the graph an inform resource
14:12:58 <ivan> chimezie: yes it is
14:13:12 <ivan> lee: sandro would you want to return to that later?
14:13:15 <ivan> sandro: yes
14:13:15 <AxelPolleres>
14:13:19 <LeeF> ISSUE-51?
14:13:19 <trackbot> ISSUE-51 -- Shall dataset clauses be allowed in SPARQL/update? -- open
14:13:19 <trackbot>
14:13:47 <ivan> Lee: how do you define a dataset for a graph pattern matching in update
14:14:00 <ivan> .... we spent a lot of time on that, including a task force
14:14:16 <ivan> ... the document has the text the tf agreed on
14:14:25 <LeeF> PROPOSED: Close ISSUE-51 based on the current design of WITH/USING/USING NAMED to specify the RDF dataset for an Update operation 
14:14:57 <ivan> +1
14:15:25 <ivan> lee: paul, from your point of view, this is stable, isn't it?
14:15:32 <ivan> pgearon: yes
14:15:42 <LeeF> RESOLVED: Close ISSUE-51 based on the current design of WITH/USING/USING NAMED to specify the RDF dataset for an Update operation 
14:15:48 <LeeF> trackbot, close ISSUE-51
14:15:48 <trackbot> ISSUE-51 Shall dataset clauses be allowed in SPARQL/update? closed
14:15:55 <LeeF> ISSUE-52?
14:15:55 <trackbot> ISSUE-52 -- Do we need the availability of an unnamed graph in SD? -- open
14:15:55 <trackbot>
14:16:01 <sandro> Lee okay to return to 49 at any point now.
14:16:21 <AndyS> Is it ever not available?
14:16:24 <ivan> lee: discussing this without greg is difficult
14:16:50 <ivan> lee: we will come back to that
14:16:59 <ivan> lee: back to issue 49,
14:17:16 <ivan> sandro: the text in the draft, I do not disagree with anything it says
14:17:17 <LeeF> PROPOSED: Close ISSUE-49 based on the current text in Section 8 at
14:17:22 <ivan> +1
14:17:34 <AndyS> 0
14:17:45 <LeeF> RESOLVED: Close ISSUE-49 based on the current text in Section 8 at, AndyS abstraining
14:17:48 <sandro> +1   the text isn't perfect   :-]     but it's good enough and not worth more debate
14:17:50 <ivan>  member:trackbot, close ISSUE-49
14:17:52 <LeeF> s/abstraining/abstaining
14:18:00 <chimezie> Sandro: if you have feedback on how to clarify the text would be welcome :)
14:18:00 <ivan> trackbot, close ISSUE-49
14:18:00 <trackbot> ISSUE-49 Is a graph an information resource closed
14:18:16 <LeeF> ISSUE-55?
14:18:16 <trackbot> ISSUE-55 -- Can/how can a separator string be supplied to the GROUP_CONCAT aggregate? -- open
14:18:16 <trackbot>
14:18:48 <ivan> lee: the decision is to add a syntax to define the extra separator character
14:18:53 <ivan> q+
14:18:56 <LeeF> PROPOSED: Close ISSUE-55 based on the use of a scalar argument SEPARATOR=";" as in
14:19:05 <LeeF> ack ivan
14:19:26 <LeeF> ivan: when I looked at the text itself, i realized that GROUP_CONCAT Is almost not defined in the document - I had to go to a mysql document to find out what it means
14:19:34 <LeeF> ... think it could use more editorial explanation
14:19:50 <ivan> lee: steve, is it on your radar
14:20:04 <ivan> steve: it needs more text
14:20:18 <ivan> ... technically it is correct
14:20:43 <ivan> +1
14:21:07 <AndyS> +1
14:21:16 <LeeF> RESOLVED: Close ISSUE-55 based on the use of a scalar argument SEPARATOR=";" as in, LeeF abstaining
14:21:17 <SteveH> +1
14:21:19 <bglimm> +1
14:21:25 <LeeF> trackbot, close ISSUE-55
14:21:25 <trackbot> ISSUE-55 Can/how can a separator string be supplied to the GROUP_CONCAT aggregate? closed
14:21:32 <LeeF> ISSUE-56?
14:21:32 <trackbot> ISSUE-56 -- Does HTTP PATCH affect either the SPARQL Protocol or the SPARQL Uniform etc. HTTP etc. Protocol? -- open
14:21:32 <trackbot>
14:21:50 <LeeF> PROPOSED: Close ISSUE-56 based on the text at
14:22:02 <ivan> q+
14:22:12 <AndyS> q+
14:22:33 <LeeF> ack ivan
14:22:35 <ivan> ivan: chime, did timbl's comment address/affect this part?
14:23:45 <LeeF> chimezie: we might want to wait - i have one open question - if you direct a PATCH at a graph IRI it suggests that you are only manipulating that graph, but SPARQL Update can affect other graphs
14:24:01 <LeeF> ... you might return a status code like Method Not Supported, but it's uncertain
14:24:11 <LeeF> ack AndyS
14:24:26 <LeeF> AndyS: What is the significance of "RECOMMENDED" in capital text?
14:24:49 <LeeF> chimezie: RFC 2119 allows RECOMMENDED to be used in plcae of SHOULD
14:25:04 <LeeF> ... i use RECOMMENDED where I did not want the same force of SHOULD
14:25:09 <ivan> chimezie: I use recommended when i did not want to use the same force than should
14:25:33 <ivan> AndyS: then i am not very happy with the text; patch is not widely used and implemented yet
14:25:51 <ivan> lee: it is also not good to use rfc terms in an informal sections
14:25:55 <SteveH> I'm not sure I agree that "few" systems support PATCH, libcurl does, and many things are based on that
14:26:07 <SteveH> q+
14:26:24 <ivan> AndyS: we should then make it a bit weaker, we should be noting that _if_ it is used, it has a particular meaning
14:26:27 <SteveH> q-
14:26:36 <ivan> ... recommending should is quite strong
14:26:52 <ivan> ... that means that is the way it should happen unless you have good reason to do something else
14:26:56 <ivan> chimezie: ok, understood
14:27:07 <LeeF> q?
14:27:10 <AndyS> q+
14:27:15 <LeeF> ack AndyS
14:27:55 <ivan> AndyS: is this discussion on sparql protocol rather than an http protocol issue
14:28:24 <ivan> chimezie: we did not reach a conclusion on that, there was then an additional comment why we did not have that
14:28:35 <ivan> ... that is the reason why it is informative
14:28:56 <ivan> AndyS: that is confusing because people will and do use PUT with sparql update requests
14:29:04 <LeeF> s/PUT/POST
14:29:24 <ivan> chimezie: a post 'here' and a post for other protocols will be different, but there is an ambiguity in general that we have to document
14:30:09 <ivan> LeeF: at some point we will have to find dedicated time to look at the protocol and define the relationships more exactly
14:30:26 <ivan> ... maybe an overview document that axel and i work on will have to clarify that
14:31:10 <LeeF> ACTION: Lee to make sure that the relationship between SPARQL protocol and HTTP protocol is clearly laid out before Last Call
14:31:10 <trackbot> Created ACTION-290 - Make sure that the relationship between SPARQL protocol and HTTP protocol is clearly laid out before Last Call [on Lee Feigenbaum - due 2010-08-17].
14:32:09 <ivan> SteveH: what was the reason we split the two documents
14:32:16 <AndyS> q-
14:32:25 <ivan> lee: we never had a strong discussion on why we have them separate
14:32:31 <pgearon> +q
14:32:34 <ivan> ... the http was a pretty new document
14:33:02 <pgearon> q-
14:33:06 <ivan> chimezie: I recall that as a major motivation, we also wanted to have a strong RESTFUL constraint 
14:33:14 <ivan> SteveH: riiiight:-(
14:33:37 <ivan> ... the protocol document with update is restful...
14:33:46 <ivan> ... it was just an idle thought
14:33:53 <ivan> lee: it is still something we can consider
14:34:05 <ivan> ... to define the overall relationships between the two
14:34:07 <ivan> ... we will consider it
14:34:22 <ivan> lee: the remaining 3 issues remain open
14:34:53 <ivan> ... 57 is close to a resolution, leaning towards the features with other but we backed off
14:35:22 <ivan> ... various implementers took actions to send descriptions to the mailing list, and I have an actions with test cases, that is not yet done
14:35:45 <ivan> ... issue 58, register mime type for sparql
14:36:03 <ivan> sandro: what we do is to add some text into the document
14:36:26 <ivan> ... from the process perspective it can go anywhere, it has to be a rec track one and that is it
14:36:30 <AndyS> Yes
14:36:35 <LeeF>
14:36:37 <ivan> lee: the query one is in the query document
14:36:38 <AxelPolleres> had two  more potential issues not sure whether we need to add them here... but also don't want to forget them
14:36:56 <ivan> lee: it is probably the update document then
14:37:09 <ivan> sandro: it is corrct
14:37:18 <ivan> s/corrct/correct/
14:37:59 <ivan> sandro: I will send a mail to paul with a possible alternative formatting
14:38:21 <ivan> pgearon: I would have thought it should go in a document that has anything to do with protocol
14:38:28 <AndyS> MIME type is related to syntax/language
14:38:28 <ivan> lee: we had it in query...
14:38:33 <sandro> Here's how I did the RIF one, following the IETF formatting more:
14:38:51 <ivan> lee: and then?
14:39:29 <ivan> sandro: once it is published (probably in LC), someone has to raise on the ietf list, and when we get to rec, there is another process to do
14:39:37 <AxelPolleres> q+ to ask about two additional (protocol related) issues
14:39:45 <ivan> lee: paul, if it is o.k. with you, I would prefer you'd put it into the update doc
14:39:52 <LeeF> ACTION: Paul to work MIME type registration information into SPARQL Update document based on Sandro's suggestion for formatting etc.
14:39:52 <trackbot> Created ACTION-291 - Work MIME type registration information into SPARQL Update document based on Sandro's suggestion for formatting etc. [on Paul Gearon - due 2010-08-17].
14:40:00 <ivan> pgearon: the more places it appears in the better is, no problems
14:40:03 <LeeF> q?
14:40:07 <LeeF> ack AxelPolleres
14:40:07 <Zakim> AxelPolleres, you wanted to ask about two additional (protocol related) issues
14:40:30 <LeeF>
14:40:44 <ivan> AxelPolleres: I had additional problems and I wonder whether they should go to issue list
14:41:01 <ivan> ... do we need a return format for update?
14:41:15 <ivan> ... if we simply get yes/no than we may be fine with the old one
14:41:20 <LeeF> ISSUE: What does the response to a SPARQL Update request look like in SPARQL protocol?
14:41:20 <trackbot> Created ISSUE-60 - What does the response to a SPARQL Update request look like in SPARQL protocol? ; please complete additional details at .
14:41:28 <AndyS> I had assumed the result body of an update would be empty.
14:41:29 <ivan> ... the other question where we would say something about transactions
14:41:34 <LeeF> AndyS++
14:41:37 <ivan> ... where do we put it
14:41:45 <ivan> ... probably the protocol document
14:42:16 <ivan> lee: my gut feeling is that the content of a response will be empty for update, but it is better to have a formal issue on that
14:42:27 <ivan> AxelPolleres: we had discussion about conveying additional infos
14:42:30 <SteveH> would like to put informative stuff in there, like number of triples added etc
14:42:32 <pgearon> http 200/400/500 are the most likely responses
14:42:55 <pgearon> oh, and 401
14:42:58 <ivan> lee: the other one probably just need the editors to take an action on documenting adomicity
14:43:03 <SteveH> 403
14:43:17 <SteveH> and 201
14:43:22 <LeeF> ACTION: Lee to make sure text on transactionality/concurrency gets added to protocol document
14:43:22 <trackbot> Created ACTION-292 - Make sure text on transactionality/concurrency gets added to protocol document [on Lee Feigenbaum - due 2010-08-17].
14:43:23 <SteveH> oh, only for PUT
14:43:27 <AndyS> any HTTP response code
14:43:44 <SteveH> +1 to AndyS 
14:44:12 <chimezie> ACTION: chimezie to modify HTTP Update document to reduce language regarding PATCH and to not use RFC 2119 language
14:44:12 <trackbot> Created ACTION-293 - Modify HTTP Update document to reduce language regarding PATCH and to not use RFC 2119 language [on Chimezie Ogbuji - due 2010-08-17].
14:44:13 <ivan> lee: the question was whether we would like to have a shortcut for update; we are not yet ready to close that
14:44:17 <AxelPolleres> can one add additional information to the success responses? e.g. if the endpoint wants to provide some additional information such as "300 triples added" ?
14:44:39 <Zakim> -SteveH
14:45:00 <ivan> Topic: document publishing status/plans
14:45:21 <ivan> Lee: how far away are we from a pretty stable document
14:45:29 <Zakim> +Garlik
14:45:30 <ivan> ... do we need one more publication before LC
14:45:37 <SteveH> Zakim, Garlik is temporarily me
14:45:37 <Zakim> +SteveH; got it
14:45:53 <ivan> ... if the document is not yet ready for that, we may need an in-depth review
14:46:13 <ivan> ... looking at editorials as well as content wise
14:46:17 <ivan> q+
14:46:42 <LeeF> ack ivan
14:47:37 <ivan> lee: query?
14:48:16 <ivan> SteveH: we need one more round before LC; there is quite a bit more content to be added
14:48:38 <ivan> AndyS: yes, I agree, it would be a good idea to have a non-lc publication
14:48:57 <ivan> ... i have been working on the property path part, that has to be moved across to the query document
14:49:24 <ivan> ... the main thing to move sections into the algebra, and this has to be quite slick before lc
14:49:43 <ivan> lee: do you want to move the pp into the document before next publication
14:49:46 <ivan> AndyS: yes
14:50:06 <SteveH> prefer to wait til after changes
14:50:21 <ivan> lee: do you want a comprehensive review now or only after things are done that you want to do
14:50:37 <ivan> AndyS: i need a really good one for the lc
14:50:53 <ivan> ... and there is only that much review people can do
14:51:03 <ivan> q+
14:51:07 <LeeF> ack ivan
14:51:59 <ivan> lee: update?
14:52:14 <AxelPolleres> I think we should have at least one new reviewer, obviously previous reviewers should also be welcome to give further comments.
14:52:25 <ivan> pgearon: a few things have to go in, that has to be done before we have a review
14:53:15 <LeeF> potential across the board approach:
14:53:26 <LeeF> 1) Editors put into documents all missing pieces
14:53:36 <LeeF> 2) Publish documents as public WD
14:53:47 <LeeF> 3) Begin in-group comprehensive reviews of all documents
14:53:56 <LeeF> 4) Incorporate review changes into Last Call-ready editors drafts
14:54:06 <LeeF> 5) Publish Last Call
14:54:14 <ivan> 6) be happy
14:54:25 <AndyS> Will next pub round [*] include tests?
14:54:38 <bglimm> 6)
14:55:05 <ivan> lee: birte, chime, are you happy with that?
14:55:23 <ivan> chimezie: i could use some community discussion before publication
14:55:24 <AxelPolleres> AndyS, I think we should include tests.
14:55:24 <bglimm> Zakim, unmute me
14:55:24 <Zakim> bglimm should no longer be muted
14:55:32 <ivan> ... eg, timbl's comment
14:55:57 <ivan> ... or do you plan to have comment incorporated before publications
14:56:36 <ivan> ... if you could distill tim's comment before next call, and see what reactions we get also on the swig list
14:56:50 <ivan> chimezie: if I can write down the issues for next week, that would be good
14:57:13 <ivan> LeeF: the wg has to have a quick turn around; ie, the answers should come from the wg
14:58:25 <ivan> --- adjourned ---