15:00:32 RRSAgent has joined #ldp 15:00:32 logging to http://www.w3.org/2013/02/11-ldp-irc 15:00:34 RRSAgent, make logs public 15:00:36 Zakim, this will be LDP 15:00:36 ok, trackbot, I see SW_LDP()10:00AM already started 15:00:37 Meeting: Linked Data Platform (LDP) Working Group Teleconference 15:00:37 Date: 11 February 2013 15:00:50 +Arnaud 15:00:51 rgarcia has joined #ldp 15:01:03 +[IPcaller] 15:01:09 +[IPcaller.a] 15:01:13 +TallTed 15:01:21 dret has joined #LDP 15:01:36 Zakim, mute me 15:01:36 TallTed should now be muted 15:01:39 Zakim, who's here? 15:01:39 On the phone I see +1.214.537.aaaa, JohnArwe, Arnaud, [IPcaller], [IPcaller.a], TallTed (muted) 15:01:39 Zakim: [IPcaller] is me 15:01:41 On IRC I see dret, rgarcia, RRSAgent, antonis, Zakim, JohnArwe, stevebattle, Ashok, TallTed, SteveS, sergio, cygri, bblfish, AndyS, bhyland, oberger, Arnaud, betehess, trackbot, 15:01:41 ... Yves, sandro, ericP 15:01:44 zakim, IPcaller is me 15:01:44 +Ashok; got it 15:01:51 +??P13 15:01:59 Zakim: [IPcaller.a] is me 15:02:03 +[IPcaller] 15:02:12 +[IBM] 15:02:13 +??P17 15:02:16 zakim, [IPcaller] is me 15:02:16 +dret; got it 15:02:18 zakim, [IBM] is me 15:02:18 +SteveS; got it 15:02:25 +SteveBattle 15:02:28 zakim, who is on the phone? 15:02:28 On the phone I see +1.214.537.aaaa, JohnArwe, Arnaud, Ashok, [IPcaller.a], TallTed (muted), ??P13, dret, SteveS, ??P17, SteveBattle 15:02:30 zakim, ??P17 is me 15:02:30 +rgarcia; got it 15:02:43 zakim, ??P13 is me 15:02:43 +AndyS; got it 15:02:45 Zakim: [IPcaller.a] is me 15:02:49 (maybe) 15:02:58 roger has joined #ldp 15:03:40 I guess [IPcaller.a] is me, but Zakim is ignoring my commands 15:03:48 zakim, who is here? 15:03:48 On the phone I see +1.214.537.aaaa, JohnArwe, Arnaud, Ashok, [IPcaller.a], TallTed (muted), AndyS, dret, SteveS, rgarcia, SteveBattle 15:03:50 On IRC I see roger, dret, rgarcia, RRSAgent, antonis, Zakim, JohnArwe, stevebattle, Ashok, TallTed, SteveS, sergio, cygri, bblfish, AndyS, bhyland, oberger, Arnaud, betehess, 15:03:50 ... trackbot, Yves, sandro, ericP 15:04:00 krp has joined #ldp 15:04:07 zakim, aaaa is Cody 15:04:07 +Cody; got it 15:04:19 +roger 15:04:41 nmihindu has joined #ldp 15:04:56 zakim: [IPcaller.a] is me 15:05:23 zakim, [IPcaller.a] is me 15:05:24 +sergio; got it 15:05:33 -TallTed 15:06:33 zakim, who's here? 15:06:35 On the phone I see Cody, JohnArwe, Arnaud, Ashok, sergio, AndyS, dret, SteveS, rgarcia, SteveBattle, roger 15:06:38 On IRC I see nmihindu, krp, roger, dret, rgarcia, RRSAgent, antonis, Zakim, JohnArwe, stevebattle, Ashok, TallTed, SteveS, sergio, cygri, bblfish, AndyS, bhyland, oberger, Arnaud, 15:06:38 ... betehess, trackbot, Yves, sandro, ericP 15:06:38 +??P12 15:07:33 I'm happy to volunteer 15:07:36 apologies; pulled into an urgent call. *may* be able to rejoin later. 15:07:49 thanks, stevebattle! 15:08:33 scribe: stevebattle 15:08:49 cody has joined #ldp 15:08:51 +??P27 15:08:53 scribenick: stevebattle 15:09:06 topic: minutes from last week 15:09:08 zakim, ??P27 is me 15:09:08 +krp; got it 15:09:14 Proposed accept minutes 15:09:21 +1 15:09:21 +1 15:09:26 +1 15:09:34 minutes approved 15:09:59 No open actions closed 15:10:09 No pending issues 15:10:13 Zakim, ??P12 is me 15:10:13 +nmihindu; got it 15:10:20 topic: issue-10 15:10:35 issue-10 15:10:35 ISSUE-10 -- Include clarifications and guidance around ETags -- open 15:10:35 http://www.w3.org/2012/ldp/track/issues/10 15:10:49 +1 15:10:51 +1 :) 15:10:51 Propose that we accept Raul's change 15:10:55 +1 15:10:55 +1 15:10:56 +1 15:10:57 +1 15:11:01 +1 15:11:05 +1 15:11:09 +1 15:11:10 +1 15:11:12 +1 15:11:56 +1 15:11:57 +Yves 15:12:01 (from agenda) link to change proposal http://lists.w3.org/Archives/Public/public-ldp-wg/2013Feb/0053.html 15:12:04 all issues can be found here: http://www.w3.org/2012/ldp/track/issues 15:12:33 +1 15:12:35 proposal carried: issue-10 closed 15:12:43 specific ones on the agenda usually are linked from the agenda, such as http://www.w3.org/2012/ldp/wiki/Meetings:Telecon2013.02.11] 15:12:49 issue-34 15:12:49 ISSUE-34 -- Adding and removing arcs in weak aggregation -- open 15:12:49 http://www.w3.org/2012/ldp/track/issues/34 15:12:53 topic: issue-34 15:13:39 +bblfish 15:13:57 Arnaud: We should use John's proposal as a reference. 3 pages with different concept is too difficult. 15:14:06 I'd like to narrow the scope. 15:14:21 Maybe we can even close issue 34? 15:14:35 q? 15:14:35 The ontology is still up for discussion 15:14:47 PROPOSED: Close ISSUE-34 Aggregation proposals selecting John's proposal as the basis moving forward. 15:14:55 +1 15:14:58 http://www.w3.org/2012/ldp/wiki/ISSUE-34 15:15:11 Direct link: http://www.w3.org/2012/ldp/wiki/Issue-34:_Back_to_Basics 15:15:17 +q 15:15:31 ack rgarcia 15:15:45 Raul: This proposal doesn't entail any chnage in the specification right now? 15:15:56 I just think that the unresolved topics will just continue in the refinement of Johns proposal, or they will spill over into issue 37 ... 15:15:58 s/chnage/change/ 15:16:04 +1 15:16:06 That's correct 15:16:06 +1 15:16:09 +1 15:16:22 +1 15:16:26 +1 15:16:27 +1 15:16:32 +1 15:16:34 +1 15:16:55 @roger: certainly some things unresolved. just in the course of doing the wiki page I found a couple (listed at bottom) and said they're follow-ons 15:17:16 0 (unfamiliar with details of the issue) 15:17:31 What is resolved? 15:17:32 ...the old tradeoff of solving all problems with it vs some useful subset + further evolution 15:17:35 Issue Resolved 15:17:36 RESOLVED: Close ISSUE-34 Aggregation proposals selecting John's proposal as the basis moving forward. 15:17:43 ah 15:17:44 ok 15:17:45 The issue is resolved 15:18:41 Ta 15:19:26 Arnaud: One of the main differences is the class hierarchy - this might be a separate issue we can open. If you have a separate proposal please submit. 15:19:46 topic : LDP Model 15:19:58 issue-37? 15:19:58 ISSUE-37 -- What is the LDP data model and the LDP interaction model? -- open 15:19:58 http://www.w3.org/2012/ldp/track/issues/37 15:20:41 Arnaud: I didn't see much discussio. EripP drafted material for the spec to introduce the model. Henry added additional material. We need to choose one rather than accumulating more text. 15:21:07 I think it's probably ok for the moment, there was a lot of discussion 15:21:47 Sorry, I don't hear much from here. 15:22:00 Arnaud: (To bblfish) what was your motivation in adding the text? 15:22:09 we hear you fine 15:22:46 bblfish: We agree that creating a container is possible - may need an issue for container creation. 15:23:05 Thanks, @John. This is very educational so far. I already have some good ideas of things to go offline and review/study after this call (throughout the week). 15:23:10 Aranud: Was this a choice between your's and EricP's text? 15:23:39 bblfish: Yes it was another proposal (we couldn't edit the original at that time). 15:23:59 s/Aranud/Arnaud/ 15:24:30 s/EricP/Eric/ 15:24:36 q? 15:24:48 s/Eric/dret/ 15:24:50 the text we're talking about is http://www.w3.org/2012/ldp/wiki/ISSUE-37#Proposed_Spec_Section_.28for_the_LDP_Spec.29 right? 15:24:58 svillata has joined #ldp 15:25:04 ok, will do. 15:25:16 yes that's the text 15:25:17 JohnArwe: right 15:25:20 topic: issue-11 15:25:21 issue-11 15:25:21 ISSUE-11 -- Do we need to define server-managed properties or do we leave them to applications? -- open 15:25:21 http://www.w3.org/2012/ldp/track/issues/11 15:25:45 +??P30 15:25:49 AndyS, i will send my comparison/combination of the two porposals to the list. this is how we should get convergence. 15:25:57 Zakim, ??P30 is me 15:25:57 +svillata; got it 15:26:10 Arnaud: Should we close or have a discussion? 15:26:16 all, if you have more comments/alternatives/criticisms, please send them to the list, and it will be my task to address them 15:26:18 PROPOSED: Close ISSUE-11 Server-managed properties accepting Raúl's proposed change. 15:26:25 q+ 15:26:33 ack bblfish 15:26:55 q+ 15:26:56 bblfish: This is about the Atom model. Is metadata in the content or the header. 15:27:24 creating a container is a different open issue 15:27:25 bblfish: I'm not sure about this anymore. 15:27:29 ack sergio 15:28:05 sergio: I'd prefer to keep both options open at this stage. 15:28:15 +EricP 15:28:33 sergio: I don't want to confuse the status of the server with the content. 15:29:08 +1 (to accept Raul's proposal) 15:29:26 q+ 15:29:31 I don't see how these could be confused. 15:30:04 q+ 15:30:07 Arnaud: The spec says the server SHOULD ignore these values. Raul wants to leave it open. 15:30:20 ack sergio 15:30:45 Arnaud said MUST, not SHOULD, ignore. Which, aside from caps, is what the spec copy/paste says in the issue body. 15:30:55 ack bblfish 15:30:57 sergio: These properties aren't intended for server data, only resource metadata. 15:31:39 +sandro 15:31:46 bblfish: Is there a requirement that these metadata properties should be synchronized with the content? 15:31:49 ok 15:31:50 and what about properties in the headers and in the content diverging? 15:32:43 bblfish: You shouldn't be able to change properties such as time? 15:33:21 SteveS: The relationship with link headers needs a new issue. 15:34:01 i think i'd even say MAY ignore, or "MAY choose to manage server-side". we should just list those that clients SHOULD be prepared to see, because they are designated as part of the LDP protocol. 15:34:15 +1 15:34:20 +1 15:34:21 +1 15:34:22 +1 15:34:22 +1 15:34:23 +1 15:34:23 It's certainly more efficient. Otherwise servers have to look at the whole content 15:34:28 +1 15:34:31 +1 15:34:32 Proposal: Close - accept Raul's proposal. The server May use these heaader properties 15:34:34 +1 15:34:37 +1 15:34:37 +0 15:34:44 +0 15:34:49 +0, bordeline -1 15:35:15 yves: What if you have properties that conflict between header and content? 15:35:39 q+ 15:35:54 ah I see 15:36:05 ack stevebattle 15:36:06 if we define the protocol to even allow conflicting location of data, then we really should say which one has preference. but in the end, maybe it's safer to simply not allow data in different locations. 15:36:33 stevebattle: it's important to be able to permit the metadata to disagree with the headers 15:36:43 s/headers/content/ 15:36:50 as dret says 15:36:55 avoid conflict, or be explicit 15:37:06 dret: We should have a clear model for which takes precedence. 15:37:14 i'd prefer to avoidn conflict. adds conmplexity and brittleness. 15:37:22 yep 15:37:22 On GET, return both? One over the other? Add a "meta" flag to request? 15:37:50 AndyS, return both what? 15:38:10 dret - property values 15:38:32 Arnaud: close issue-11 15:38:38 q+ 15:38:42 close issue-11 15:38:42 Closed ISSUE-11 Do we need to define server-managed properties or do we leave them to applications?. 15:38:48 q? 15:38:51 zakim, pick an issue 15:38:51 I don't understand 'pick an issue', dret 15:39:21 Issue-36 15:39:21 ISSUE-36 -- Can applications create new containers? -- open 15:39:21 http://www.w3.org/2012/ldp/track/issues/36 15:39:25 bblfish: what about issue 36 15:39:28 +1 to issue-36 15:39:30 topic issue-36 15:39:56 Arnaud: I believe the answer is Yes - but doesn't specify how it is done. 15:40:04 q? 15:40:26 ack bblfish 15:40:29 Arnaud: Just create a resource with the right type. 15:40:55 bblfish: Sounds reasonable. Use a POST with content of type container, or content with link type Container? 15:41:26 i'd prefer POSTing a container representation that contains all the metadata that's required to create a container 15:41:45 Arnaud: This header link type isn't in the spec right now? 15:41:49 -sergio 15:41:55 SteveS: confirms 15:41:59 q? 15:42:03 q+ 15:42:10 I prefer POSTing a container representation, just like any other resource type 15:42:11 ack sergio 15:42:46 sergio: The issue is - Do you want to introspect the data when creating the container? 15:43:58 q+ 15:44:07 Arnaud: 1) The spec allows it job done 2) Make it explicit in the spec 15:44:19 ? 15:44:22 what's the implementation impact of "data introspection"? you have to process and validate the incoming data against protocol errors anyway, so spotting some simple pattern to dispatch to "container creation" is not such a big deal. 15:44:33 AndyS: It all works until you want to create a container in a container. 15:45:08 ack steves 15:45:08 AndyS: We're approaching a solution; this is just about the details. 15:45:23 i think we shouldn't constrain conatiners to always being hierarchical. many things on the web are sets/bags of things, and not trees. 15:46:11 SteveS: I'm thinking along the lines of no special mechanism. It's the same as for resources. Do we need to say anything - it might be helpful to introduce informative text to explain how containers are 'born'. 15:46:15 i think we should have examples around container creation. people will want to know. 15:46:30 +1 to what SteveS said 15:46:39 +1 15:46:48 yes, we should clarify. It's a cool feature 15:47:10 Aranud: SHould we close, requesting editors to explain the birds & the bees? 15:47:28 s/Aranud/Arnaud/ 15:47:56 q? 15:47:59 Arnaud: Do you want to make a proposal (Henry)? (about linke headers) 15:48:07 s/linke/link/ 15:48:36 PROPOSED: close ISSUE-36, leaving to the editors to propose some text clarifying how this is done in the spec 15:48:43 +1 15:48:53 +1 15:48:55 +1 15:48:58 +1 15:49:12 +1 15:49:13 +1 15:49:15 +[IPcaller] 15:49:17 +1 15:49:20 +1 15:49:21 +1 15:49:23 +1 15:49:26 +1 15:49:47 close issue-36 15:49:47 Closed ISSUE-36 Can applications create new containers?. 15:50:11 SteveS: What about issue-33 15:50:17 topic issue-33 15:50:22 topic: issue-33 15:51:39 Arnaud: People wanted containers for the pagination mechanism. Should we move pagination to LPD Resource (Containers would inherit it). 15:52:11 paging is always "page by", right? how would that work when you cannot "page by member"? 15:52:36 SteveS: Simply change restriction on the predicate used. 15:52:39 @dret, are you asking "what are the units of paging"? 15:52:58 yes, AnyS. hard to imagine what to even page by. 15:53:08 s/Anys/AndyS/ 15:53:22 q+ 15:53:39 by "triple" seems the obvious answer 15:53:40 Arnaud: The history is amusing because we had pagination for all resources in the original. We felt it needed simplification. 15:53:43 ack sergio 15:54:01 sergio: This rule breaks general linked data principles. 15:54:20 your factoring it out had the desired effect of making LDBP easier for people to understand and use as a starting place 15:54:22 q+ 15:54:29 ack stevebattle 15:54:32 paging often is a URI Template driven process with: page size, page number, or it is driven jyst by opaque URIs for prev and next. what to page by is almost always implicit. how is that specified in the generlized model? 15:54:32 s/it out/pagination out/ 15:54:33 q+ 15:55:12 in the spec it is opaque; just next (and optionally prev) links 15:55:14 Agree: this does not seem to break Linked Data principles 15:55:15 stevebattle: The Linked Data API supports pagination. 15:55:34 What do you page by? 15:55:43 LDA does not paginate graphs - it paginates by the (list) results 15:55:47 ack steves 15:56:00 LDA is not only graphs out. 15:56:01 Arnaud: Page on the entire content of the resource, not a specific predicate. 15:56:17 you just page by triples in a pretty much random bag of triples? 15:56:38 I think there is a pageSize URI query parm, that's where units would arise. 15:56:46 SteveS: The server is often best placed to know how to page the content. 15:57:25 i'd qualify arnaud's explaination with the caveat is possible when there are some logical partitions within the data. even if it's triples with a predicate starting with a, and b, ... 15:57:34 Arnaud: The first and subsequent pages have different URIs so this is consistent with the model. 15:57:34 dret, does that make sense to you? 15:57:59 how is that going to be stable over time, when the triples are more or less randomly arranged? i must be missing something here? 15:58:06 I am happy with a general mechanism 15:58:26 PROPOSED: close ISSUE-33: adding pagination for non-container resources 15:58:33 +1 15:58:35 ericP, i see a lot of tight coupling ot brittleness in there, but i still think i must be missing something. 15:58:36 +1 15:58:37 +1 15:58:39 roger has joined #ldp 15:58:41 +1 15:58:42 +0 15:58:50 +0 15:59:01 +1 15:59:06 +1 15:59:07 and ot course there's always HTTP, if you really just want to serve partitions of bigger resources. HTTP does support this natively... 15:59:08 @dret, "stable" paging was server impln choice. 15:59:11 (unclear on paginating graphs - triples, entities, random?) 15:59:11 +1 15:59:13 +1 15:59:13 +1 15:59:16 -1 15:59:20 same than AndyS 15:59:26 +1 15:59:28 sorry, too much to decide on that right now. -1 15:59:31 q+ 15:59:41 q+ to propose that the next step is a use case 15:59:47 dret: I don't understand how this is supposed to work. 15:59:57 ack ericp 15:59:57 ericP, you wanted to propose that the next step is a use case 16:00:19 +0 (kind of agree with Eric. need to see how this is useful in a scenario) 16:00:24 EricP: The next step should be a use-case, to ensure that we undertstand how to use pagination in a non-container. 16:00:30 -[IPcaller] 16:00:38 q- 16:00:40 if it's just serving big data, we really should look at HTTP: it can do this out of the box. 16:00:40 ok 16:00:47 If a server or app can't do paging, it don't need to 16:00:51 thanks 16:00:54 thanks 16:00:54 adjourned 16:00:55 -roger 16:00:57 -sandro 16:00:58 -bblfish 16:00:58 -EricP 16:00:59 thanks everybody! 16:00:59 -Yves 16:00:59 -Ashok 16:00:59 s/it don't/it doesn't/ 16:01:00 -Cody 16:01:02 meeting adjourned 16:01:03 -krp 16:01:04 -rgarcia 16:01:04 -svillata 16:01:05 -dret 16:01:05 -SteveS 16:01:06 -AndyS 16:01:07 Thanks Arnaud, bye 16:01:08 -JohnArwe 16:01:11 -Arnaud 16:01:15 -SteveBattle 16:01:16 -nmihindu 16:01:18 SW_LDP()10:00AM has ended 16:01:18 Attendees were +1.214.537.aaaa, JohnArwe, Arnaud, TallTed, Ashok, dret, SteveS, SteveBattle, rgarcia, AndyS, Cody, roger, sergio, krp, nmihindu, Yves, bblfish, svillata, EricP, 16:01:18 ... sandro, [IPcaller] 16:01:47 SteveS_ has joined #ldp 16:02:25 SteveS__ has joined #ldp 16:45:26 SteveS_ has joined #ldp 16:45:51 SteveS_ has joined #ldp 17:48:02 TallTed has joined #ldp 18:15:16 jmvanel has joined #ldp 18:29:30 Zakim has left #ldp 18:40:44 cygri has joined #ldp 19:37:49 SteveS has joined #ldp 19:38:06 SteveS has joined #ldp 20:32:08 jmvanel has joined #ldp 20:44:47 SteveS has joined #ldp 20:45:05 SteveS has joined #ldp 21:44:50 SteveS_ has joined #ldp 22:48:15 bhyland has joined #ldp 23:05:55 bblfish has joined #ldp