14:58:07 RRSAgent has joined #ldp 14:58:07 logging to http://www.w3.org/2014/11/17-ldp-irc 14:58:09 RRSAgent, make logs public 14:58:09 Zakim has joined #ldp 14:58:11 Zakim, this will be LDP 14:58:11 ok, trackbot; I see SW_LDP()10:00AM scheduled to start in 2 minutes 14:58:12 Meeting: Linked Data Platform (LDP) Working Group Teleconference 14:58:12 Date: 17 November 2014 14:58:20 azaroth has joined #ldp 14:59:21 SW_LDP()10:00AM has now started 14:59:28 +azaroth 14:59:47 +Alexandre 15:00:01 +Arnaud 15:00:14 bblfish has joined #ldp 15:00:16 +Sandro 15:00:26 +EricP 15:01:25 Ashok has joined #ldp 15:01:46 +deiu 15:02:05 +Ashok_Malhotra 15:02:14 -Alexandre 15:02:41 +Alexandre 15:02:57 +[IBM] 15:03:09 zakim, [IBM] is me 15:03:09 +SteveS; got it 15:04:39 + +33.6.47.14.aaaa 15:05:26 zalim, aaaa is me 15:05:32 zakim, aaaa is me 15:05:33 +bblfish; got it 15:05:51 MiguelAraCo has joined #ldp 15:05:55 ScribeNick: azaroth 15:06:04 TOPIC: Minutes of last meeting 15:06:18 http://www.w3.org/2013/meeting/ldp/2014-11-10 15:06:20 Arnaud: I suggest we approve them, objections? Hearing none, minutes of Nov 10 are approved 15:06:35 ... Next meeting will be next week, Nov 24. 15:06:47 ... Nothing new in tracker, one action on LD Patch. 15:06:49 https://dvcs.w3.org/hg/ldpwg/raw-file/ldpatch/ldpatch.html#grammar-production-add 15:07:02 Alexandre: Pasted in a link to the page I just pushed 15:07:22 ... Page about Patch. 15:07:34 Arnaud: Close the action, and discuss later in the call 15:07:37 action-150? 15:07:37 action-150 -- Alexandre Bertails to Add support for arbitrary text/turtle for the add operation in ldpatch -- due 2014-10-06 -- OPEN 15:07:37 http://www.w3.org/2012/ldp/track/actions/150 15:07:46 TOPIC: Status of the specs 15:07:56 Arnaud: Eric, do we have the call yet? 15:07:58 +OpenLink_Software 15:08:06 Zakim, OpenLink_Software is temporarily me 15:08:06 +TallTed; got it 15:08:08 Zakim, mute me 15:08:08 TallTed should now be muted 15:08:09 Eric: Pinging Ralph on IRC now, not set up yet 15:08:33 Arnaud: Need to get this done soon. 15:08:46 ... Sent a request to move paging to CR 15:08:52 ... SHould be able to have a single call for both 15:09:12 ... Please lets not delay it any further. Out of charter at the end of the month, would like to be PR by then 15:09:33 ... Happy at the same time to see updated reports coming in 15:09:44 ... Makes the story better. Thank you Steve for keeping up with them 15:10:03 ... Looking at paging, what's the status of the tests? 15:10:07 ... Anyone working on it? 15:10:24 ... Have a draft ready, with pointer to suite, but more a framework and we have to develop the tests 15:10:35 ... I know Steve has been working on some, but is anyone contributing? 15:11:06 Steve: The bodies of the tests haven't been written, just a skeleton for the framework 15:11:14 ... No pull requests to incorporate anything 15:11:17 https://dvcs.w3.org/hg/ldpwg/raw-file/tip/tests/ldp-paging-testsuite.html 15:11:24 ... Unless someone with commit rights has done it themselves 15:11:36 ... Waiting for people to provide implementations on individual tests 15:11:52 Arnaud: Getting an error when I click on the link (above) 15:12:08 Alexandre: I think you shouldn't be using tip, use default 15:12:12 https://dvcs.w3.org/hg/ldpwg/raw-file/default/tests/ldp-paging-testsuite.html 15:12:50 Arnaud: We have a stub but who is planning to contribute tests, anyone? 15:13:00 Steve: I plan to, but dont have a solid plan 15:13:08 ... Maybe towards end of year or early next year 15:13:13 Arnaud: Anyone else? 15:13:21 ... Henry? 15:13:23 basically in Mercurial, tip points to the latest commit, across *all* branches, so a push to another branch will break it. Prefer a named branch itself: much more stable 15:13:25 *tumbleweed* 15:13:35 I am not implementing paging yet. 15:13:35 I/we are 15:13:35 Arnaud: Who is implementing paging 15:13:42 *more tumbleweed* 15:13:49 I need to have a JavaScript client first then paging becomes important 15:14:00 Arnaud: Not completely surprised. Sandro has been open that he's not interested 15:14:18 q+ 15:14:21 ... Not clear that we have enough interest to get to PR 15:14:40 ... need to have people implement, and get tests 15:14:40 ack bblfish 15:15:02 Henry: There's a notion of creating a special serialization of RDF, then binary paging could work 15:15:18 ... paging in HTTP, but need to organize the RDF in a specific way. Would also help for patch. 15:15:33 ... interested in paging later 15:15:38 not yet sorry 15:15:44 Arnaud: So not interested as currently defined 15:15:52 or maybe you wont, because you'll see it's not useful. 15:15:57 ... Spec may remain in CR for the foreseeable future. has happened before 15:16:06 Sandro: We're not supposed to do that if the WG ends 15:16:26 ... Used to do it, but not supposed to leave things open on REC track 15:16:37 Arnaud: Can turn it into a note later 15:16:48 Sandro: Unless we're not rechartered 15:17:05 Arnaud: Assuming we're going to recharter, yes. Details to be defined but assume there'll be another WG 15:17:13 ... that's it for the status of specs 15:17:17 TOPIC: LDPatch 15:17:39 Arnaud: Proposal to publish as CR. Talked about it last week. A lack of feedback. 15:17:40 -1 to passing LD Patch to LC from me :-) 15:18:02 ... Have a stable spec, got some public comments. Not clear what's stopping us, but no one prepared to make the first move 15:18:13 ... Can consider it, would like to hear what people thing 15:18:13 q+ 15:18:16 s/thing/think 15:18:18 s/thing/think/ 15:18:31 ... Can move it to CR, but if not good enough, what's wrong? 15:18:32 Here is the issue tracking wiki page: https://www.w3.org/2012/ldp/wiki/Ldpatch-issues 15:18:37 ack betehess 15:19:02 Alexandre: Changes from ?? commons have a big impact, and need to be reflected in the intro 15:19:15 ... too many todos to publish now, but in a week we could reconsider 15:19:24 no on eis proposing to publish as-is 15:19:25 Arnaud: Update with where you are with draft? 15:20:03 [[ 15:20:04 * section 2: LD Patch s/format/semantics/ 15:20:04 * make it normative 15:20:04 * extract examples into a 3rd non-normative section 15:20:04 * add HTTP semantics section with content of current 5. 15:20:04 * explain UpdateList semantics with drawing 15:20:06 * remove Abstract Syntax section 15:20:07 Alexandre: Tim's comment simplifies some things. Have remaining thing to do ... address comments from Steve 15:20:08 * move 2.3 Pathological graph at the end of introduction 15:20:08 where is current draft? 15:20:10 ]] 15:20:17 q+ to ask what timeframe are we looking at for publication? 15:20:23 ... once we've done those things we'll be happy. Can be done by end of week 15:20:26 ack me 15:20:27 ack deiu 15:20:27 deiu, you wanted to ask what timeframe are we looking at for publication? 15:20:44 deiu: We talked about LC for patch, but not when. Is there a deadline? 15:21:03 Arnaud: Don't have a well defined plan. Need to see what's missing 15:21:29 ... progress from editors, not clear what was going to happen. Happy to give until next week 15:21:40 deiu: Some changes have been done, e.g. the grammar 15:21:47 ... Now mostly editorial changes 15:21:58 ... Once these are done it should be ready for review 15:22:23 Arnaud: Move to CR/LC as we have the shortened process. Cna choose to combine the two 15:22:37 ... Can take advantage of that. Any comments? 15:22:58 ... invite people to comment and implement, and do disposition of comments at the same time 15:23:02 ... assume everyone is okay? 15:23:26 ... Trying to figure out how to get to rec faster. 15:23:35 ... WG have been given the option to combine them 15:23:45 ... Unclear if this actually helps 15:23:55 ... As have to do all the steps, just last two at the same time 15:24:07 ... If there's changes you have to go back to CR anyway 15:24:12 ... Questions? Comments? 15:24:16 ... Also need a test suite 15:24:28 Alexandre: Have a test suite, finalized yesterday by pierre antoine 15:24:43 Arnaud: Give the editors a week to finish and then freeze to review for a week? 15:24:56 Sandro: I think I heard it could be done by firday, could review on the weekend? 15:25:05 The spec is not that long btw 15:25:08 Arnaud: Not safe to ask people to review on the weekend 15:25:22 Steve: Travelling on the weekend and have a short week next week 15:25:28 ... could try to review by Dec 1 15:25:45 ... Would have to go through the changelog, which ones have and haven't been handled 15:25:55 ... would be nice to see on the wiki 15:26:22 Alexandre: Biggest was tim's comment, which was some refactoring 15:26:28 Arnaud: Any chance to have it done by friday? 15:26:33 Alexandre: I can try 15:26:43 Sandro: If Steve can't review by Dec 1, does it matter? 15:26:48 Arnaud: Steve? 15:26:56 Steve: Depends when, have a hard stop at 2pm 15:27:35 Arnaud: As soon as it's done send an email to the group. If early enough, great, otherwise we'll have to wait 15:27:49 Steve: Should we ask how many people plan to support/implement? 15:27:56 I am 15:27:58 Arnaud: Who is going to implement LD patch 15:28:09 Alexandre: We have two implementations 15:28:12 For me it's allready implemented since Alex's implementation is in Scala :-) 15:28:13 We are planning to do some patch work 15:28:25 q+ 15:28:32 Steve: Planning to do it for some things 15:28:36 ack bblfish 15:28:55 Henry: If Alexandre doing in scala, I don't need to implement :) but also a possible JS implementation 15:29:03 Arnaud: Anyone else? 15:29:04 due to scala-js.org 15:29:22 Sandro: One of the students working for Tim will probably end up implementing it 15:29:30 I think Tim already started implementing it 15:29:32 Arnaud: I think that's it for LD patch 15:29:37 Ie. if alex does the implementation in Scala it can be compiled to JavaScript. 15:29:38 ... let's talk about rechartering 15:29:44 betehess, JS for use in NodeJS 15:29:54 ... Sandro sent out a long list of possible features 15:30:03 ... some are familiar, others are not at all 15:30:21 ... clear from the discussion that it resonated with some people 15:30:29 sandro, should work in node.js and in browsers 15:30:44 ... up to us what we want to put in the charter. took a first shot at it 15:30:55 ... just a strawperson for the WG to start working on 15:31:20 ... would like to hear from people based on discussion what things we can all agree on 15:31:47 ... also talked about time frame. From discussion, there's agreement we don't want to rush into working on the next version 15:32:02 ... I think there's a way to set up the schedule so we get the first 6 months as exploratory period 15:32:10 ... would allow us to refine what we want to do 15:32:20 ... suggestion is not to worry too much about that 15:32:29 ... when we have a list of things we want to do, we can update the timeline 15:32:48 ... lets focus on what we want to tackle 15:33:07 ... Sandro, you put up that list and got feedback. Do you have a sense what might be of interest to the larger group? 15:33:14 Sandro: I've completely forgotten :) 15:33:36 Sandro’s email: http://lists.w3.org/Archives/Public/public-ldp-wg/2014Nov/0010.html 15:33:40 what's the link to the draft charter? 15:33:49 https://www.w3.org/2012/ldp/wiki/LDPNext2015_Charter 15:33:59 ... Haven't read the emails since last week, don't remember. Did look over the draft charter. Could expand the list a bit. 15:34:28 ... Don't really define the scope very much. thoughts about rewriting the introduction. RDF + REST ... not sure how well it has aged 15:34:35 Arnaud: Anyone? 15:34:44 I think it works very well the RDF+REST 15:34:59 ... Clearly some of the REST community laugh at this. At the same time, it still conveys the right message? 15:35:04 Steve: What would you replace it with? 15:35:06 I don't see anything about Queries in the deliverable? 15:35:08 Sandro: Haven't gotten that far 15:35:22 why are you cringing? 15:35:24 Steve: I had a similar reaction, but thought it hit the right nerve as well 15:35:25 RDF+JSON-LD+REST ? 15:35:35 ah, I understand, the first deliverable defers to Scope 15:35:40 ... trying to leverage HTTP and rich data model 15:35:58 ? 15:36:00 Sandro: Independence, modularity, kind of thing 15:36:13 don't understand what " Independence, modularity, kind of thing" ? 15:36:19 Ted?: Hypertect should lead to all future states? 15:36:36 s/Ted?/ericP/ 15:36:53 Sandro: that we should modularize http. from REST perspective shouldn't bundle 20 functions into one thing, just add independent features. SO sort of violated rest principles 15:36:55 mhh, look at HTTP spec? 15:37:02 ericP: WOrking examples that we could look at? 15:37:14 Arnaud: Steve posted a blog post on how to create a container 15:37:18 ... got some reaction 15:37:29 I agree we could have the spec simpler without LDP diect and indirect containers, that could go into a seperate spec. But otherwise... 15:37:32 ... reflection of the sentiment of REST folk that LDP is not restful 15:37:35 what is that blog post? 15:37:41 I think LDP *is* very RESTful 15:37:53 Steve: Don't feel like adding more to the claims. tightly coupled. 15:38:00 SteveS: can you point us to the blog 15:38:05 Sandro: Took me several emails to understand. Then i got it and agreed 15:38:06 what are you guys referring to??? 15:38:23 ericP: Nice goal, or require rearchitecture? 15:38:25 ( thanks betehess, I have the same question ) 15:38:41 Steve: Didn't see it as a binary thing. 15:38:49 ... loosely coupled with tighter semantics 15:39:08 q+ 15:39:13 Sandro: Some things end up more tightly coupled than they need to be. If loose coupling had been a goal, we would have made different design decisions 15:39:18 Steve: Not really sure what it means 15:39:21 Henry: URL? 15:39:24 ack bblfish 15:39:27 Steve: A bunch of conversations 15:39:43 ... some reaction to creating a container. POST some RDF and headers. 15:39:54 A god a twitter conversation! :D 15:39:58 Arnaud: Reaction was on twitter, not really even a conversation 15:40:03 ... more like a sarcastic comment 15:40:03 I am dying of laughter here 15:40:06 q- 15:40:23 ... not sure how to make progress with draft charter 15:40:26 q+ 15:40:40 ... don't think we can take on everything from Sandro's mail 15:40:43 ... just a handful 15:41:08 ... would like people to have a look. scope has turned into a list of questions. could be different. 15:41:17 ... by next week, would be good if people could consider the list 15:41:35 q+ 15:41:35 ... questions that should not be there, ones that should and are not 15:41:47 Sandro: By email, rather than by wiki? 15:42:10 Arnaud: lets go thru existing list 15:42:18 ... then email after that 15:42:20 https://www.w3.org/2012/ldp/wiki/LDPNext2015_Charter#scope 15:42:45 ... same framework as the current charter 15:43:11 ... q1. "How can retrieval of a container and its contained resources be combined so that fewer HTTP operations are required than it is necessary with LDP 1.0?" 15:43:23 ... what we've called inlining resources 15:43:34 ... people still interested? 15:43:43 +1 15:43:44 yes, sounds interesting. Requires N-Quads 15:43:46 +1 from Stanford on this 15:43:47 we at Pellucid already do Inlining 15:43:49 +1 15:44:10 Ashok?: We have an option where you can ask for inlining 15:44:11 not sure what to standardize though 15:44:31 Sandro: WOnder a little about framing it differently 15:44:40 ... if we have query support you might not talk about it the same way 15:44:47 (me either nquads or a type of literal "<...> <...> <...>"^^lang:Turtle . 15:44:52 Arnaud: Querying not there, I believe. Should we add it? 15:45:00 Sandro: From my perspective yes 15:45:01 yes +1 for adding querying. 15:45:14 Arnaud, Sandro: 3 sounds like a query 15:45:18 q+ 15:45:20 -1 to 3 being query 15:45:33 are we using the queue? 15:45:46 Sandro: Simple filtering mechanism could subsume 1 15:45:50 +1 to looking at both inline and query (which results might blend) 15:45:51 Steve: not sure 15:45:56 query can be used to meet 1 and 3 but 3 doesn't intrinsically subsume 1 15:46:05 ok, even the chair ignores the queue :-) 15:46:06 +1 to betehess and the use of the queue 15:46:17 +0.5 for query, not as high for me 15:46:48 (some discussion) 15:47:26 Arnaud: request filtering. WIth 1 and 3 I think we cover the ground. 15:47:33 -EricP 15:47:35 ... What about q2 about multiple resources in a single request 15:47:44 Sandro: I would generalize to also include updated 15:47:48 +1 for 2 with POST on Tar 15:47:48 +EricP 15:47:53 ... might take out HTTP 15:48:04 ... does websockets count as HTTP? 15:48:16 Arnaud: I don't know 15:48:17 +1 for bulk operations 15:48:21 ... anyone else? 15:48:29 q? 15:48:37 Henry: We should use the queue 15:48:49 ack bblfish 15:49:07 ... to go back, query is one thing and orthogonal is 1. That's GET where get contained resources. 15:49:36 ... second one is creating multiple resources, there's HTTP2/SPDY but also how do you create resources by posting a Tar for example 15:49:43 ack betehess 15:50:05 Alexandre: we have already a query language that covers some of the points 15:50:12 ... do filtering, aggregation, inlining etc 15:50:34 ... was a workshop discussed? do we want to do it? 15:50:38 Sandro: Not a bad idea 15:51:04 ... could try to reach out to new people, having a workshop on standard data query languages or narrower 15:51:20 Arnaud: concern would be how much of that we want to do as part of the group 15:51:28 ... hope we don't get into defining a query language 15:51:36 Sandro: That's why we haven't done it so far 15:51:51 Alexandre: we published a blog post a few weeks ago, most would be open source 15:52:06 ... main thing is we want other people to work on the same thing 15:52:17 ... would need to become members, then could do a submission 15:52:26 I think there is something about the need for a query langauge for LDP, so again either narrow SPARQL down to LDP or a new language 15:52:34 Sandro: CG could be a reasonable place. Maybe do that and bring them into the WG if they come up with something 15:52:37 ... like JSON-LD 15:52:39 Alexandre, can you post a ptr? 15:53:07 Arnaud: Some kind of workshop interesting. Like a WG meeting, but open as invite the world. needs travel, so added cost 15:53:29 Modifying my answer: query could be a bit chunk of work, it would be good to understand in more detail what is broken with current approaches (SPARQL) and seen some work that has shown promise 15:53:29 Sandro: WGs do stakeholder meetings. WG meeting, then days of public meeting with others who have some connection, like use cases or adopters 15:53:34 here is the pointer of our work on what we internally call LDQL: http://io.pellucid.com/blog/exposing-resources-in-datomic-using-linked-data 15:53:39 Arnaud: Timing of that ? 15:53:50 Sandro: Within the next 6 months 15:54:06 ... good to colocate it? Could do a couple of events at related conferences... 15:54:14 +1 to colocate it 15:54:22 Arnaud: How do we charter to accomodate work in that space? 15:54:30 Sandro: Already kind of doing it, given the questions 15:54:55 ... could be more specific ... is there a better way to extract something by a query information about the container and its members 15:55:11 -1 to query language definition 15:55:23 +1 15:55:25 +1 to 4 15:55:29 Arnaud: What about q4 -- how can a client be notified? 15:55:35 Ashok: This is Sandro's push ? 15:55:36 +1 to 4 15:55:37 4. is interesting but I don't see it part of LDP 15:55:51 Sandro: Not just a resource, but anything, eg query results 15:56:12 agree that this is query based 15:56:14 ... you have a container of millions of resources and want to know what happens in the container, sounds like a query 15:56:14 PubSubHub? RSS feed? 15:56:16 +1 to 4, work out either/both push or pull model 15:56:39 +1 to 4, in conjunction with SocialWeb WG activity streams 15:56:51 mhh. how do you ask that something has changed? 15:56:56 q+ 15:57:14 ack bblfish 15:57:31 Henry: You can do in sparql, seems like something like sparql update. asking to query a change 15:57:44 Sandro: 3 different research implementations trying to do it with sparql standing queries 15:57:48 SPARQL Triggers. 15:57:57 yay, reinventing SQL! 15:57:59 ... C-Sparql might be most well known, continuous sparql 15:58:19 Arnaud: q5 15:58:23 TallTed: not re-inventing, just improving by starting with more solid foundations :-) 15:58:31 +1 for adding yet another link header ;) 15:58:45 ... don't want to do it for every resource, so need a mechanism to do for a set of resources 15:58:56 Sandro: Generalize beyond sparql, might want to do it for description mechanism 15:59:25 ... link header doesn't work, could be end point that handles 10M resources, don't want to access them all 15:59:45 ... something like .well-known 15:59:54 +1 to sandro, for generalizing it for things like SPARQL and event mechanism 16:00:11 Arnaud: Out of time. should focus on that. Will send updated list to mailing list, can comment on it 16:00:21 ... so a week from now we have a better list 16:00:30 ok, got to take train to Paris from Cannes now ... 16:00:30 ... thanks all, thanks Rob for scribing, talk to you next week 16:00:31 -Sandro 16:00:32 -Ashok_Malhotra 16:00:32 -SteveS 16:00:33 -Alexandre 16:00:35 bye 16:00:35 -Arnaud 16:00:37 -TallTed 16:00:41 -bblfish 16:00:42 -azaroth 16:00:44 -deiu 16:00:57 -EricP 16:00:57 SW_LDP()10:00AM has ended 16:00:57 Attendees were azaroth, Alexandre, Arnaud, Sandro, EricP, deiu, Ashok_Malhotra, SteveS, +33.6.47.14.aaaa, bblfish, TallTed 16:03:00 sandro has left #ldp 16:03:14 bblfish has joined #ldp 17:21:26 SteveS has joined #ldp 17:26:17 SteveS has joined #ldp 18:22:26 Zakim has left #ldp