14:58:16 <RRSAgent> RRSAgent has joined #ldp
14:58:16 <RRSAgent> logging to http://www.w3.org/2014/12/01-ldp-irc
14:58:18 <trackbot> RRSAgent, make logs public
14:58:18 <Zakim> Zakim has joined #ldp
14:58:20 <trackbot> Zakim, this will be LDP
14:58:20 <Zakim> ok, trackbot; I see SW_LDP()10:00AM scheduled to start in 2 minutes
14:58:21 <trackbot> Meeting: Linked Data Platform (LDP) Working Group Teleconference
14:58:21 <trackbot> Date: 01 December 2014
14:58:47 <SteveS> SteveS has joined #ldp
14:59:12 <pchampin> pchampin has joined #ldp
15:00:32 <Zakim> SW_LDP()10:00AM has now started
15:00:39 <Zakim> +Arnaud
15:00:58 <Ashok> Ashok has joined #ldp
15:01:05 <Zakim> +bblfish
15:01:12 <Zakim> +deiu
15:01:23 <Zakim> +OpenLink_Software
15:01:29 <TallTed> Zakim, OpenLink_Software is temporarily me
15:01:29 <Zakim> +TallTed; got it
15:01:31 <TallTed> Zakim, mute me
15:01:31 <Zakim> TallTed should now be muted
15:01:37 <Zakim> +Sandro
15:01:40 <bblfish> hi
15:01:52 <Zakim> +Ashok_Malhotra
15:01:55 <Zakim> +[IBM]
15:01:56 <Zakim> +??P14
15:01:57 <codyburleson> codyburleson has joined #ldp
15:02:03 <SteveS> zakim, [ibm] is me
15:02:03 <Zakim> +SteveS; got it
15:02:12 <pchampin> zakim, ??P14 is me
15:02:12 <Zakim> +pchampin; got it
15:02:35 <Zakim> +[IPcaller]
15:02:50 <codyburleson> Zakim, IPcaller is me
15:02:50 <Zakim> +codyburleson; got it
15:03:48 <bblfish> yes, I could scribe
15:03:56 <bblfish> but I may fall off due to bad connection
15:04:46 <bblfish> Topic: Aproval of minutes
15:04:51 <Arnaud> Proposal: Approve the minutes of the 24 November teleconf: http://www.w3.org/2013/meeting/ldp/2014-11-24
15:05:02 <Arnaud> Resolved: Approve the minutes of the 24 November teleconf: http://www.w3.org/2013/meeting/ldp/2014-11-24
15:05:21 <bblfish> will meet next time December 8
15:05:38 <bblfish> ( I will be at Scala eXchange in London, so probably won't be able to make it )
15:05:46 <Zakim> +Alexandre
15:05:58 <Arnaud> PROPOSED: Open ISSUE-102: Negative indexes in Slice-s
15:06:00 <SteveS> +1
15:06:02 <betehess> +1
15:06:04 <pchampin> +1
15:06:09 <bblfish> +1
15:06:10 <TallTed> +1
15:06:18 <Arnaud> RESOLVED: Open ISSUE-102: Negative indexes in Slice-s
15:06:24 <Arnaud> PROPOSED: Open ISSUE-103: Hard vs silent fail on missing delete triples
15:06:29 <SteveS> +1
15:06:29 <pchampin> +1
15:06:30 <betehess> +1
15:06:32 <TallTed> +1
15:06:34 <codyburleson> 0
15:06:36 <bblfish> +1
15:06:38 <deiu> +1
15:06:43 <Arnaud> RESOLVED: Open ISSUE-103: Hard vs silent fail on missing delete triples
15:07:06 <Zakim> +ericP
15:10:46 <bblfish> ah sorry
15:10:58 <SteveS> ISSUE-102?
15:10:58 <trackbot> ISSUE-102 -- Negative indexes in Slice-s -- raised
15:10:58 <trackbot> http://www.w3.org/2012/ldp/track/issues/102
15:10:58 <bblfish> what's the topic?
15:11:32 <bblfish> Topic: LDPatch issues
15:12:22 <betehess> q+
15:12:52 <bblfish> deiu: repeating the position on the issue
15:13:22 <bblfish> It broke up a bit for me here
15:13:25 <MiguelAraCo> MiguelAraCo has joined #ldp
15:13:28 <Arnaud> ack betehess
15:13:30 <bblfish> I don't think I am hearing that well
15:14:15 <bblfish> It's not going to be that good for me to take notes here
15:14:21 <bblfish> I can't hear the arguments
15:14:54 <pchampin> betehess: Tim's remark about list indexes is two-fold
15:15:19 <pchampin> ... 1/ editorial remark, suggesting we describe list operations at a higher level
15:15:30 <pchampin> ... without relying on rdf:first and rdf:rest
15:15:46 <Arnaud> ack sandro
15:16:06 <pchampin> ... 2/ Tim suggested we allowed negative indexes as they are used in Python
15:16:19 <pchampin> ... to denote elements from the end of the list
15:16:46 <pchampin> ... which we think is bad, as the implementation would have to browse the entire list to know its size
15:17:07 <TallTed> Zakim, who's noisy?
15:17:14 <pchampin> sandro: this should be done, the use case where I want to insert elements to the end of a list
15:17:17 <Zakim> TallTed, listening for 10 seconds I heard sound from the following: ericP (59%), Arnaud (9%), deiu (4%)
15:17:24 <TallTed> Zakim, mute ericP
15:17:24 <Zakim> ericP should now be muted
15:17:28 <pchampin> ... without having to know the size of the list, is a very useful one
15:17:46 <betehess> my remark is more like that: there is a little _computational_ cost, but nothing I am afraid of, especially compared to what it can bring, as Sandro said :-)
15:17:48 <pchampin> ... there is no additional cost (although scribe didn't really get the argument here)
15:18:11 <Arnaud> PROPOSED: Close ISSUE-102 adding Negative indexes in Slice-s
15:18:12 <pchampin> q+
15:18:25 <Arnaud> ack pchampin
15:18:32 <betehess> +0.99
15:19:31 <SteveS> +1
15:19:37 <deiu> +1
15:19:38 <ericP> +1
15:19:40 <Ashok> +1
15:19:44 <ericP> sandro: +1
15:19:48 <pchampin> pchampin: if we add them in slices, it raises the question of adding them in paths as well
15:19:55 <TallTed> +0.75
15:20:13 <codyburleson> 0
15:20:18 <TallTed> Zakim, unmute me
15:20:18 <Zakim> TallTed should no longer be muted
15:20:20 <pchampin> pchampin: +1
15:20:31 <TallTed> Zakim, mute me
15:20:31 <Zakim> TallTed should now be muted
15:20:34 <Arnaud> RESOLVED: Close ISSUE-102 adding Negative indexes in Slice-s
15:20:53 <Arnaud> ISSUE-103: Hard vs silent fail on missing delete triples
15:20:53 <trackbot> Notes added to ISSUE-103 Hard vs silent fail on missing delete triples.
15:21:33 <pchampin> betehess: in the current spec, Delete fails if asked to delete non-existing triples
15:22:29 <Ashok> q+
15:22:32 <pchampin> ... others would prefer that Delete silently pass on non-existing triples
15:22:37 <SteveS> q+
15:22:45 <pchampin> ... as the intention is simply that the triple be absent in the resulting graph
15:23:10 <pchampin> ... and that stronger constraints should be handled with ETags for example
15:23:54 <pchampin> ashok: why would this be an error to delete a non-existing triple?
15:24:23 <pchampin> betehess: Tim wants to be able to detect if someone changed the resource in between
15:24:51 <deiu> q+
15:24:59 <pchampin> ... and does not want to use etags for this, but rather optimistic concurrency
15:25:22 <Arnaud> ack Ashok
15:25:26 <pchampin> ashok: would it be ok to simply raise a warning?
15:25:49 <pchampin> sandro: I would not be ok, because he does not want the change to happen
15:25:53 <Arnaud> ack SteveS
15:26:04 <pchampin> betehess: if we start with warnings, we would need a way to "rollback"
15:26:46 <pchampin> steves: why focus on delete; part of the discussion was also about Add, not failing when asked to add an *existing* triple
15:26:50 <Arnaud> ack deiu
15:27:18 <pchampin> betehess: that's true, but Tim's remark was about Delete only
15:27:20 <betehess> I think that solving the issue for Delete will provide a similar solution for Add
15:27:43 <Arnaud> ack sandro
15:27:44 <pchampin> deiu: ETags do not make sense when several people are changing different parts of the same resource
15:29:20 <pchampin> sandro: if you have to change independantly parts of a resource, you should probably have several resources
15:29:41 <Zakim> -bblfish
15:30:13 <SteveS> I like to say, I put in my patch the desired end state of my resource…so operations will not be needed to achieve that
15:30:18 <pchampin> arnaud: may be we should not express this in terms of silent fail
15:30:24 <SteveS> s/so operations/some operations/
15:30:29 <TallTed> Zakim, unmute me
15:30:29 <Zakim> TallTed should no longer be muted
15:30:29 <ericP> note that SQL has syntax switchs for this
15:30:32 <betehess> solutions: a) 2 modes are supported b) delete on missing triple fails c) delete on missing triple does not fail
15:30:32 <pchampin> ... but in terms of what the LD Patch is asking
15:31:12 <pchampin> ... if Delete aks that the triple is not there in the resultin graph,
15:31:20 <pchampin> ... then it is a success if the triple is already not there
15:31:21 <betehess> q+
15:31:30 <pchampin> ... and symetric for Add
15:31:42 <Arnaud> ack betehess
15:31:49 <pchampin> ... So one possibility is to have two variants of Add and Delete
15:32:00 <Zakim> +bblfish
15:32:20 <pchampin> ... Another one is to have Assertions (as proposed by pchampin on the mailing list)
15:32:22 <TallTed> Zakim, mute me
15:32:22 <Zakim> TallTed should now be muted
15:32:35 <pchampin> betehess: a third way is to have a globale mode for the patch, lax or strict
15:32:52 <betehess> s/glogale/global/
15:33:25 <SteveS> having a global mode and then perhaps per operation to override/guarantee something, like Delete(mode=[lax | strict])
15:35:37 <pchampin> pchampin: assertions gives more flexibility, but may force you to be more verbose
15:37:00 <pchampin> sandro: I have no strong opinion on etags / strict mode / assertions
15:37:14 <betehess> so far: a) assertions b) mode strict/lax c) 2 sets of operations d) etags + delete doesn't fail
15:37:36 <pchampin> ... what should you do in strict mode ? completely rollback if one operation fails?
15:37:53 <Ashok> yes
15:37:54 <pchampin> steves: yes, that's the semantics of the PATCH verb
15:38:04 <pchampin> sandro: that's a heavy burden
15:38:50 <TallTed> Zakim, unmute me
15:38:50 <Zakim> TallTed should no longer be muted
15:40:26 <pchampin> pchampin: we could constrain assertions to be in the preamble, before any operation
15:40:28 <betehess> I don't understand the database argument
15:40:31 <ericP> why wouldn't it be an issue?
15:40:45 <pchampin> tallted: we are again down the path of reinventing database
15:40:47 <Ashok> You would have transactions
15:40:55 <Zakim> -bblfish
15:40:58 <ericP> just because the database can support the desired behavior doesn't mean we have language to signal the desired behavior
15:41:11 <betehess> I don't understand how transactions are relevant here, we are only updating one resource, not several at once
15:43:29 <Zakim> +bblfish
15:43:39 <TallTed> no argument, ericP -- I'd be happy to have the language for signaling this... but that implies a larger burden on servers to support handling those signals, and on clients to send them
15:43:46 <betehess> q+
15:43:56 <Arnaud> ack betehess
15:44:08 <pchampin> sandro: why not just rely on etags now, and add another feature in a further version?
15:44:42 <pchampin> betehess: we still have to decide on the default behaviour, which will have to be conservative in further versions
15:45:06 <pchampin> arnaud: sandro's point is that the default would be Delete fails silently
15:45:16 <pchampin> ... this minimal option is appealing
15:45:40 <Arnaud> STRAWPOLL: a) assertions b) mode strict/lax c) 2 sets of operations d) etags + delete doesn't fail
15:45:55 <pchampin> q+
15:45:58 <SteveS> 0.0 0.2 0.1 1.0
15:46:03 <Arnaud> ack pchampin
15:46:29 <betehess> -0   +1   +0.5   +0.5
15:47:00 <deiu> 0 +1 0 +1
15:47:12 <Ashok> b or c
15:47:59 <pchampin> 1/3 1/3 0 1/3
15:48:04 <TallTed> true neutral, at this point...
15:48:06 <codyburleson> a) 0 b) 0 c) 0 d) +0.5
15:49:25 <pchampin> tallted: 2 sets of operations can be useful to chose the operations only *when* you want to be sure
15:49:51 <pchampin> pchampin: @tallted yes I can't see why you would care only on a part of the patch
15:51:09 <pchampin> arnaud: we are running out of time; let people think about it until next week
15:51:39 <pchampin> topic: working group rechartering
15:51:46 <betehess> also, timbl brought up an issue on the error status code... but he didn't send that on the public mailing list (did he?)
15:52:57 <pchampin> arnaud: we argued that we have work in progress ; mentionned the CSS WG as an example
15:53:18 <pchampin> ... but it does not seem to be an example to follow
15:53:34 <pchampin> ... the easiest would be to ask for an extension
15:54:31 <pchampin> ... and we already have some positive feedback from W3M
15:55:02 <betehess> can't we ask like 3 months at a time? could avoid asking for too much
15:55:10 <pchampin> sandro: IMO we could ask 6m or 1y
15:56:37 <pchampin> arnaud: @betehess, it is not really good to keep asking for an extension
15:56:50 <pchampin> ... better to ask once, and when we are ready propose a new charter
15:57:31 <pchampin> ashok: we can ask for 12m, and tell if we are done earlier
15:57:43 <pchampin> arnaud: do people think that 6m would be enough?
15:57:48 <ericP> i'm with Sandro on this; go for a year
15:57:55 <betehess> in my case, 6 months is enough to do more than playing, including LD Patch
15:57:58 <SteveS> seems a little short to me, I would think end of summer 2015
15:58:21 <betehess> and I like having shorter deadlines, eg. 6 months, because you try harder to deliver things
15:58:40 <bblfish> I am happy with 6 months, but you may as well get more if you can.
15:58:53 <pchampin> sandro: we can argue that we need less resources during the extension
15:58:55 <SteveS> I agree to try to achieve before, so 6 would be a good forcing timeframe
15:59:06 <pchampin> arnaud: we could indeed reduce the frequency of calls
15:59:42 <betehess> 6 months to deliver, 3 months to wrap things up :-)
16:00:50 <betehess> thanks
16:00:51 <Zakim> -bblfish
16:00:53 <bblfish> thanks
16:00:56 <Zakim> -deiu
16:00:57 <Zakim> -Alexandre
16:00:57 <Zakim> -Sandro
16:00:59 <Zakim> -SteveS
16:00:59 <Zakim> -codyburleson
16:01:00 <Zakim> -Arnaud
16:01:00 <Zakim> -ericP
16:01:02 <Zakim> -TallTed
16:01:03 <bblfish> sorry bad connection here
16:01:07 <Zakim> -Ashok_Malhotra
16:04:43 <pchampin> scribe: pchampin
16:17:22 <Arnaud1> Arnaud1 has joined #ldp
16:20:06 <bblfish> bblfish has joined #ldp
16:26:40 <Arnaud> Arnaud has joined #ldp
16:35:01 <Zakim> disconnecting the lone participant, pchampin, in SW_LDP()10:00AM
16:35:02 <Zakim> SW_LDP()10:00AM has ended
16:35:02 <Zakim> Attendees were Arnaud, bblfish, deiu, TallTed, Sandro, Ashok_Malhotra, SteveS, pchampin, codyburleson, Alexandre, ericP
17:22:33 <Arnaud> Arnaud has joined #ldp
17:44:28 <Arnaud> Arnaud has joined #ldp
17:47:00 <SteveS> SteveS has joined #ldp
18:03:03 <Arnaud1> Arnaud1 has joined #ldp
18:03:35 <Zakim> Zakim has left #ldp
18:44:51 <Arnaud> Arnaud has joined #ldp
19:17:09 <SteveS> SteveS has joined #ldp
19:23:42 <Arnaud1> Arnaud1 has joined #ldp
19:40:45 <Arnaud> Arnaud has joined #ldp
19:43:48 <Arnaud1> Arnaud1 has joined #ldp
19:52:41 <Arnaud> Arnaud has joined #ldp
19:54:10 <deiu> deiu has joined #ldp
20:43:41 <Arnaud1> Arnaud1 has joined #ldp
21:17:39 <deiu> deiu has joined #ldp
21:23:43 <bblfish> bblfish has joined #ldp
21:59:20 <bblfish> bblfish has joined #ldp