Warning:
This wiki has been archived and is now read-only.
Chatlog 2013-03-11
From Linked Data Platform
See original RRSAgent log or preview nicely formatted version.
Please justify/explain all edits to this page, in your "edit summary" text.
14:03:52 <RRSAgent> RRSAgent has joined #ldp 14:03:52 <RRSAgent> logging to http://www.w3.org/2013/03/11-ldp-irc 14:03:54 <trackbot> RRSAgent, make logs public 14:03:56 <trackbot> Zakim, this will be LDP 14:03:56 <Zakim> ok, trackbot, I see SW_LDP()10:00AM already started 14:03:57 <trackbot> Meeting: Linked Data Platform (LDP) Working Group Teleconference 14:03:57 <trackbot> Date: 11 March 2013 14:04:35 <Zakim> -ericP 14:04:59 <Zakim> +[GVoice] 14:05:19 <Arnaud> zakim, who's on here? 14:05:20 <Zakim> I don't understand your question, Arnaud. 14:05:46 <Arnaud> zakim, who is here? 14:05:46 <Zakim> On the phone I see Sandro, Arnaud, [GVoice], pchampin, Ashok_Malhotra, ??P10, TallTed (muted), ??P13, dret, JohnArwe 14:05:48 <Zakim> On IRC I see RRSAgent, Arnaud, JohnArwe, dret, Zakim, Ashok, TallTed, pchampin, jmvanel, gavinc, sandro, ericP, Yves, thschee, betehess, trackbot 14:06:14 <ericP> Zakim, [GVoice] is me 14:06:14 <Zakim> +ericP; got it 14:08:38 <dret> i can step in 14:08:46 <dret> ok, i won't step in 14:08:53 <sandro> scribe: sandro <sandro> chair: Arnaud 14:11:00 <sandro> two unidentified callers. mysterious! 14:11:23 <dret> +1 14:11:24 <ericP> +1 <sandro>Topic: Minutes of last meeting 14:11:32 <sandro> PROPOSED: accept minutes from last week 14:11:45 <sandro> RESOLVED: accept minutes from last week (March 4) <sandro>Topic: Next meetings 14:11:27 <sandro> Arnaud: At our last F2F, we arranged for an informal telecon afterwords. 14:12:20 <sandro> Arnaud: At informal telecon right afterwords, we went over what happened, to catch up people who missed it. That's the plan for next week's telecon. 14:12:44 <sandro> Arnaud: So people who come to the F2F, you don't have to join the telecon next week. We wont be making any decisions or anything like that. 14:12:58 <sandro> Arnaud: On the F2F -- do we have the final list of participatns? 14:13:29 <sandro> Arnaud: I haven't heard any feedback on the agenda. 14:14:25 <sandro> Arnaud: Maybe breakouts. Long list of issues; no proposal for some of them. 14:15:00 <sandro> Arnaud: I STRONGLY suggest if you care about an issue, you send out a proposal BEFORE the meeting. 14:15:05 <dret> the break-out groups would be chartered with preparing proposals for those issues which don't have any right now. 14:15:20 <Ashok> q+ 14:15:51 <sandro> Arnaud: If people are remote and really care, we can try to work around their schedule. 14:16:05 <Arnaud> ack ashok 14:16:47 <sandro> Ashok: We still have arguments and disagreements on the overall @@. I was hoping we could talk about that early 14:16:55 <sandro> s/@@/model/ 14:17:20 <sandro> Arnaud: I'm kind of torn on that. 14:17:57 <sandro> Arnaud: If we could agree on the model, things would be easier, yes. But it doesn't seem to be working, so maybe bottom up will work better. 14:18:03 <Zakim> +Yves 14:18:06 <sandro> .. like develop the test suite, and see how that goes. 14:18:11 <dret> switch from topd-down to bottom-up mode, yes. i think we have established that top-down doesn't seem to work all that great for us. 14:18:29 <sandro> +1 bottom up 14:18:51 <sandro> Arnaud: We'll end up needing to make the same design decisions, but without the details we don't understand each other. 14:18:54 <dret> and then when we have better spec'd bottom-up cases, we need to distill the model. 14:19:23 <sandro> eric: There's a possibility that working with test cases will get us toward the model than discussion the model theoretically 14:19:28 <sandro> Arnaud: exactly. 14:19:35 <dret> yup, i agree that the tests are an excellent way to go forward. 14:19:52 <sandro> ericP: Ashok, do you see a way to have the Test discussion first, in a way that will help us have a grounded Model discussion. 14:19:55 <Arnaud> q? 14:20:04 <sandro> dret, can you take over scribing for 10-15 now? 14:20:12 <sandro> Ashok: I need to think about this. 14:21:34 <dret> i can do that... <dret>scribe: dret 14:22:46 <dret> Arnaud: keep breakout sessions as an option, but in such a fragmented group it may be counter-productive; will be decided on demand 14:23:06 <Zakim> -??P13 14:23:20 <dret> Topic: Actions and Issues 14:23:49 <Zakim> -??P10 14:24:13 <dret> Proposal: close ACTION-29 14:24:54 <dret> Resolved: ACTION-29 closed 14:26:22 <dret> ACTION-38 remains open while we switch to bottom-up mode for the model discussion <sandro>Topic: Open Issues <sandro>subtopic: Issue-49 14:30:00 <Ashok> But is this a LDP issue? 14:30:44 <Arnaud> ack Yves 14:31:47 <dret> discussing ISSUE-49 and whether it's part of LDP or should be considered orthogonal 14:32:46 <dret> Yves: pointing to metalink and saying that we should be very cautious about making mirroring/equivalence a part of LDP 14:32:52 <TallTed> TallTed has joined #ldp 14:33:02 <dret> Ashok: agrees with Yves 14:33:19 <dret> Arnaud: seems like most people agree that it's not an LDP issue 14:33:20 <JohnArwe> ...yves said he would put reference to other materials on the public list; some issues with attack vectors. 14:33:43 <Arnaud> ack sandro 14:34:26 <dret> Sandro: use OWL sameas in data or in the HTTP header 14:35:04 <dret> these are RDF solutions, we might also recommend HTTP-level ones? 14:35:14 <pchampin> q+ to notice that owl:sameAs is symetrical 14:35:21 <Arnaud> ack pchampin 14:35:21 <Zakim> pchampin, you wanted to notice that owl:sameAs is symetrical 14:36:28 <dret> pchampin: not sure about sameas, because it is supposed to be symmetrical 14:37:05 <SteveS> SteveS has joined #ldp 14:37:39 <Zakim> +[IBM] 14:37:39 <Yves> See https://tools.ietf.org/html/rfc6249 14:37:44 <sandro> scribe: sandro 14:38:03 <SteveS> zakim, [IBM] is me 14:38:03 <Zakim> +SteveS; got it 14:38:12 <Ashok> q+ 14:38:15 <dret> thanks, sandro! 14:38:23 <Arnaud> proposal: close issue-49, making no change 14:38:34 <Arnaud> ack ashok 14:38:59 <sandro> Ashok: There's a section 4.1.4 that spoke to this. Does that get clarified or removed or what? 14:39:16 <sandro> Arnaud: No one has proposed that 14:39:54 <sandro> Arnaud: There is one MUST in the spec. 14:40:32 <sandro> Ashok: I recommend keeping it open and figuring out what to do with that part of our spec. 14:41:04 <sandro> Arnaud: I don't want to just leave it open, and then we'll come back in 2 months 14:41:13 <sandro> Ashok: give me an action 14:41:38 <SteveS> q+ 14:41:38 <sandro> Arnaud: So we'll leave ISSUE-49 open for now, and Ashok with produce a recommendation about what to do about 4.1.4 14:41:47 <Arnaud> ack steves 14:42:14 <bhyland> bhyland has joined #ldp 14:43:42 <sandro> action Ashok to propose how to modify the text of 4.1.4 to go with closing ISSUE-49 (with no material modification to spec) 14:43:42 <trackbot> Created ACTION-40 - Propose how to modify the text of 4.1.4 to go with closing ISSUE-49 (with no material modification to spec) [on Ashok Malhotra - due 2013-03-18]. 14:44:42 <dret> i'm fine with ignoring ISSUE-37 for now and then attempt to distill it from tests <sandro>subtopic: Issue-21 14:45:22 <sandro> Arnaud: We were going to talk about Reverse Membership Predicate, but Steve Battle asked we not decide in his absence, so we'll put it off. 14:45:35 <sandro> Arnaud: He did send email with example. 14:46:59 <sandro> dret: I liked Steve's notion that navigability and data model are different things. What are the concepts that we identify and represent; and THEN decide on the affordances, where do we encode it, so clients can find that part of the data model in that place. 14:47:35 <sandro> .. I really like the idea of splitting them. Make membership part of model, then later talk about how to make it navigable, maybe in both directions. 14:47:55 <sandro> Arnaud: I thought you might agree with that. :-) 14:48:10 <sandro> Arnaud: Steve's example was very useful. I had not understood it. 14:48:20 <sandro> Arnaud: We need more examples posted. 14:49:06 <sandro> Arnaud: I had thought it was a back-link, but he's just using a different membership property. The link is in the same direciton, container to member. 14:50:13 <sandro> subtopic: issue-35 14:50:16 <sandro> issue-35? 14:50:16 <trackbot> ISSUE-35 -- POSTing to a container MUST yield a fresh URI -- open 14:50:16 <trackbot> http://www.w3.org/2012/ldp/track/issues/35 14:50:41 <sandro> Arnaud: We had a similar discussion about whether URIs could be reused after delete. 14:51:07 <sandro> .. Like MUST you get an error doing a GET after a DELETE? 14:51:22 <sandro> .. people seemed to think there may be cases where a URI is reused. 14:51:26 <sandro> issue-24? 14:51:26 <trackbot> ISSUE-24 -- Should DELETED resources remain deleted? -- closed 14:51:26 <trackbot> http://www.w3.org/2012/ldp/track/issues/24 14:51:39 <dret> i hope we don't say anywhere that you MUST get an error.. 14:51:45 <sandro> Arnaud, so issue-35 and issue-24 seem linked 14:51:59 <pchampin> I agree that they seem linked 14:52:01 <sandro> (your logic sounds good to me, Arnaud) 14:52:31 <TallTed> seems an accurate assessment... 14:52:57 <sandro> dret, I don't think we have liberty to decide this. if a server decides to re-use URIs, that's up to it. clients shouldn't make those assumptions. 14:53:12 <sandro> s/dret,/dret:/ 14:53:29 <sandro> dret: I don't think clients have business looking at URIs like this. 14:53:43 <sandro> Arnaud: Richard proposed server must not re-assign URI. 14:54:00 <sandro> q+ 14:54:19 <TallTed> POSTing to a container MUST yield a URI ... it's server's choice what that URI is, and I think whether that URI is guaranteed new or re-used is and should remain outside our control 14:54:47 <sandro> Arnaud: To be consistent with 24 we can't change anything on 35. because we already said URIs may be reused 14:54:51 <Arnaud> ack sandro 14:55:32 <ericP> sandro: i'm concearned about the situation where client A deletes </foo> and client B continues to use </foo> 14:56:03 <ericP> ... if something else creates </foo>, the meaning will change out from under client B. 14:56:23 <sandro> Arnaud: Yeah, "under certain cirumstances". So in the NORMAL case you'd get a new URI. 14:56:28 <ericP> ... we need text which talks about "delete if if you're *sure* the URL hasn't leaked out" 14:56:33 <sandro> Arnaud: So I'm okay with a SHOULD. 14:56:34 <Arnaud> ack tallted 14:56:43 <dret> i think this is "deployment guide" stuff: don't do it, because of this and that. 14:56:57 <sandro> TallTed: This is no different from any other situation where URIs change. 14:57:42 <pchampin> I agree with TallTed 14:57:49 <dret> me too 14:57:54 <sandro> TallTed: I don't think this is an undetectable error condition. You just have to document your server behavior 14:58:20 <dret> i'd prefer non-spec level, this is just best practice 14:58:57 <sandro> sandro: documenting server bevavior means there's no spec here. 14:59:18 <SteveS> I like increasing the awareness from what we have, like with should or recommended, and possibly update some guidance into the deployment guide 14:59:45 <sandro> sandro: MUST is too hard to implemnt, but strongly worded SHOULD is okay 15:00:04 <sandro> Arnaud: Sounds like we're coming to SHOULD, but let's not decide until we have Richard present. 15:00:11 <dret> not sure. this is not a protocol issue. it's just a recommendation. 15:00:57 <dret> yup, and if people want to shoot themselves in the foot, they are allowed to do so. they just dimish the vaue of their service. 15:00:59 <sandro> TallTed: I think SHOULD might be too strong. eg in the case of a reboot 15:01:19 <sandro> sandro: if the client can't assume the URIs will be stable on a given server, we have a huge problem. 15:01:21 <Zakim> -Ashok_Malhotra 15:01:23 <dret> thanks everybody! thanks, sandro! 15:01:29 <Zakim> -dret 15:01:33 <sandro> thanks, dret! 15:01:41 <sandro> Arnaud: See many of you in Boston!! 15:01:45 <sandro> ADJOURN 15:01:49 <Zakim> -Yves 15:01:50 <Zakim> -TallTed 15:02:09 <Zakim> -pchampin 15:03:17 <Zakim> -JohnArwe 15:04:34 <Zakim> -SteveS 15:05:07 <Zakim> -Sandro 15:05:08 <Zakim> -Arnaud 15:05:12 <Zakim> -ericP 15:05:13 <Zakim> SW_LDP()10:00AM has ended 15:05:13 <Zakim> Attendees were pchampin, Ashok_Malhotra, TallTed, dret, JohnArwe, Sandro, Arnaud, ericP, Yves, SteveS <sandro> present: pchampin, Ashok_Malhotra, TallTed, dret, JohnArwe, Sandro, Arnaud, ericP, Yves, SteveS # SPECIAL MARKER FOR CHATSYNC. DO NOT EDIT THIS LINE OR BELOW. SRCLINESUSED=00000229