IRC log of ldp on 2014-12-01

Timestamps are in UTC.

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
  翻译：