13:57:20 <RRSAgent> RRSAgent has joined #ldp
13:57:20 <RRSAgent> logging to http://www.w3.org/2014/08/18-ldp-irc
13:57:22 <trackbot> RRSAgent, make logs public
13:57:22 <Zakim> Zakim has joined #ldp
13:57:24 <trackbot> Zakim, this will be LDP
13:57:24 <Zakim> ok, trackbot; I see SW_LDP()10:00AM scheduled to start in 3 minutes
13:57:25 <trackbot> Meeting: Linked Data Platform (LDP) Working Group Teleconference
13:57:25 <trackbot> Date: 18 August 2014
13:57:48 <Ashok> Ashok has joined #ldp
13:59:05 <Zakim> SW_LDP()10:00AM has now started
13:59:12 <Zakim> +??P5
13:59:18 <pchampin> zakim, ??P5 is me
13:59:18 <Zakim> +pchampin; got it
13:59:52 <Zakim> +Arnaud
14:00:28 <Zakim> +Ashok_Malhotra
14:00:41 <Zakim> +Steve_Speicher
14:01:00 <SteveS> Zakim, Steve_Speicher is SteveS
14:01:00 <Zakim> +SteveS; got it
14:01:02 <Zakim> + +1.857.928.aaaa
14:01:11 <betehess> Zakim, aaaa is Alexandre
14:01:13 <Zakim> +Alexandre; got it
14:03:10 <JohnArwe> JohnArwe has joined #ldp
14:03:33 <Arnaud> zakim, who's on the phone?
14:03:33 <Zakim> On the phone I see pchampin, Arnaud, Ashok_Malhotra, SteveS, Alexandre
14:03:37 <Zakim> + +1.857.204.aabb
14:03:42 <ericP> Zakim, aabb is me
14:03:42 <Zakim> +ericP; got it
14:03:51 <JohnArwe> zakim not working so well
14:03:57 <Zakim> +Sandro
14:03:59 <Zakim> +JohnArwe
14:06:48 <betehess> scribe: Alexandre
14:06:56 <betehess> scribenick: betehess
14:07:05 <betehess> topic: minutes of last week
14:07:22 <betehess> Arnaud: can we approve them?
14:07:24 <sandro> scribe: sandro
14:07:26 <sandro> scribenick: sandro
14:07:38 <sandro> arnaud: Minutes of last week?    Some issues around paging, but we'll get back to that
14:07:50 <sandro> Arnaud: minutes approved.  meet next week?
14:08:02 <sandro> subtopic: actions and issues
14:08:16 <Zakim> +OpenLink_Software
14:08:23 <TallTed> Zakim, OpenLink_Software is temporarily me
14:08:23 <Zakim> +TallTed; got it
14:08:24 <sandro> arnaud: sandro sent email re action-145
14:08:26 <TallTed> Zakim, mute me
14:08:26 <Zakim> TallTed should now be muted
14:08:36 <sandro> arn: So who is doing that work?
14:09:00 <sandro> arn: Someone needs to make the json-ld context -- mixed with the namespace document.
14:09:23 <sandro> JohnArwe: Nandana said he could do it, according to minutes from last week
14:09:28 <sandro> Arnaud: great.
14:09:29 <JohnArwe> regrets: cody
14:10:11 <sandro> topic: paging
14:10:43 <sandro> Arnaud: Last week we decided to go to Last Call, making a few decisions.   Then JohnArwe came back from his trip and pointed out some problems with the decisions we made.
14:11:36 <ericP> Arnaud, btw, i think i screwed up 149 (meant to move it to pending review). plh sent http://www.w3.org/mid/1408370806.2900.13.camel@chacal to the apps WG chair
14:11:53 <sandro> Arnaud: Regarding units, we had number-of-triples and we decided to add more units.
14:12:20 <sandro> Arnaud: John said I dont know how to specify multiple units
14:12:35 <sandro> Arnaud: MUST or MAY -- does the server have to support this
14:13:19 <sandro> JohnArwe: How intentfully do you take the first
14:13:34 <sandro> ericP: Someone specified both number of bytes and number of triples
14:13:52 <sandro> eric: I though 7240 was talking about lexically first
14:14:13 <sandro> JohnArwe: Two consecutive headers
14:14:27 <sandro> Arnaud: Could be client just specifies one.
14:14:53 <sandro> eric: I thought use case was machine that was both memory bound and wanted to display ten units to user
14:15:17 <sandro> eric: Whichever is smallest
14:15:30 <sandro> eric: Not whether we allow people to specify multiple units, but is server responsible for respecting multiple limits
14:15:49 <sandro> JohnArwe: Also, what syntax the client uses, since the existing syntax doesn't obviously support it.
14:16:00 <sandro> q+
14:16:23 <sandro> Arnaud: Did people expect multiple units at the same time?
14:17:10 <sandro> strawpoll: 1=we need multiple syntaxes at the same time, 0=client just picks one
14:17:23 <SteveS> I don’t see need to support multiple (and server honors whichever is smallest), seems to be handling an edge case and adds more complication than needed.  Be good to ensure we could support ti.
14:17:37 <sandro> JohnArwe: It sounds like it came up on the fly, without existing use cases
14:18:21 <JohnArwe> on the strawpoll, it's mult units not mult syntaxes that's the first order question
14:18:23 <sandro> Arnaud: I suggest caution
14:18:41 <Arnaud> ack sandro
14:20:09 <sandro> sandro: Why are we using return=representation, instead of our own preference
14:21:22 <sandro> JohnArwe: We could do multiple preferences, or just have our current one with an extended syntax
14:21:41 <JohnArwe> options in play for units are triples, bytes, members
14:21:51 <sandro> JohnArwe: we just have a problem with repeating the same parameter
14:22:17 <sandro> JohnArwe: But it's just a should
14:22:22 <betehess> somebody has an example of things that could be repeated?
14:22:36 <JohnArwe> ...a 2119 should, as in "unless you have Good Reason to do otherwise"
14:23:03 <sandro> ericP: It seems like when you're parsing you have three slots, bytes, triples, records.
14:23:12 <Ashok> q+
14:23:22 <Zakim> -pchampin
14:23:26 <sandro> JohnArwe: Should use the first, should discard the others
14:23:49 <sandro> ericP: THat would be good if we're talking about the first/others on bytes, fr instance
14:23:53 <Ashok> q-
14:23:58 <Zakim> +??P5
14:24:04 <pchampin> zakim, ??P5 is me
14:24:04 <Zakim> +pchampin; got it
14:24:24 <sandro> JohnArwe:   max-triple-count="500"     (one option)
14:26:44 <JohnArwe> ...sandro intends for the sugg above to replace pagesize="500 triples"
14:26:51 <sandro> PROPOSED: let's use max-triple-count, max-record-count, max-kbyte-count  with parms being integers
14:26:53 <Ashok> +1 to Sandro's proposal
14:27:14 <Arnaud> proposed: allow multiple units using separate parameters
14:27:15 <ericP> +1
14:27:22 <TallTed> +1
14:27:24 <sandro> +1
14:27:25 <betehess> +0
14:27:28 <JohnArwe> +0
14:27:41 <SteveS> +0 fine with just triple count
14:27:56 <sandro> RESOLVED: let's use max-triple-count, max-record-count, max-kbyte-count  with parms being integers
14:29:23 <sandro> sandro: isnt there an overarching constraint that there can never be a MUST on the server on http PREFER
14:29:42 <sandro> JohnArwe: I think we could say say every LDP server must support them.
14:30:25 <TallTed> Zakim, unmute me
14:30:25 <Zakim> TallTed should no longer be muted
14:30:47 <sandro> PROPOSED: Make the three units be MUST, AT RISK for server
14:31:16 <sandro> JohnArwe: TallTed, can your code do this?  How much implementation effort is this?
14:31:30 <sandro> TallTed: Byte-Range ?
14:31:48 <sandro> JohnArwe: No one knows how to do that with RDF
14:32:04 <sandro> TallTed: Clearly the right thing for the client is to allow byte limits
14:33:18 <sandro> sandro: Given the AT RISK, we can take it out if people don't implement it.
14:33:32 <sandro> SteveS: When all three are used.... what?
14:34:42 <sandro> sandro: I picture it as the server has three different checks of "have I reached the max-triple-count, if so then close up"    "have I reached the number of bytes, then if so close up", etc
14:35:12 <JohnArwe> ...sandro, you also said that At Risk also lets us change the Musts (understand each unit, now keyword) to Mays
14:35:41 <sandro> Arnaud: it's not easy
14:35:47 <sandro> sandro: its's easy
14:35:48 <sandro> :-)
14:35:56 <sandro> PROPOSED: Make the three units be MUST, AT RISK for server
14:35:57 <betehess> +0
14:36:01 <Ashok> +1
14:36:02 <ericP> +.3
14:36:17 <sandro> +0.791
14:36:19 <JohnArwe> I completely disagree that it's easy, with Jena and especially with built-in serializers.
14:37:38 <sandro> PROPOSED: We'll including in the paging spec that we have AT RISK the three limits, first reached, and we might take out one, two, or all both before PR, or make them MAY instead of MUST.
14:37:54 <ericP> +1
14:37:58 <TallTed> +1
14:38:03 <sandro> +1
14:38:10 <pchampin> +0
14:38:16 <betehess>  /me concurs with JohnArwe re: Jena and other serializers
14:38:17 <betehess> +0
14:38:32 <SteveS> -0 think it should be a should, I was fine with just triples
14:38:32 <JohnArwe> +0 ok since at risk
14:38:43 <sandro> RESOLVED: We'll including in the paging spec that we have AT RISK the three limits, first reached, and we might take out one, two, or all both before PR, or make them MAY instead of MUST.
14:39:10 <sandro> JohnArwe: I think we're done, yeah.
14:39:15 <sandro> topic: LD-PATCH
14:39:32 <sandro> Arnaud: We have a proposal, we had some pushback from Sandro and Eric with counter-proposals
14:40:02 <Zakim> +EricP.a
14:40:08 <Zakim> -ericP
14:41:12 <sandro> Arnaud: we need to settle which direction we're taking, once and for all.
14:42:27 <ericP> sandro: i think it works but i'm skeptical that it will be adopted.
14:42:50 <ericP> ... i'm happy to see an FPWD after resolving the syntactical issues
14:43:23 <JohnArwe> hmmm... sounds like same argument (in reverse) we just used to add mult units into a LC draft for paging
14:43:28 <betehess> I personally agree on only one change s/-/^/g
14:43:55 <betehess> all the rest is different from SPARQL, that's why I don't want to make it look like it
14:44:20 <sandro> ericP: We should ask the community, and the best way to do that is to publish.     so we can't decide before.
14:44:33 <betehess> q+
14:45:08 <JohnArwe> is the assumption that the patch syntax has to fit w/in LDP's charter extension period, which IIRC is YE2014?
14:45:25 <Arnaud> ack betehess
14:45:42 <sandro> sandro: I'm not convinced the semantics are diff from sparql
14:46:02 <sandro> betehess: If the semantics are different the syntax sould be different
14:46:07 <sandro> sandro: Sure, of course.
14:46:38 <pchampin> q+
14:46:46 <betehess> oh right, the slice operator :-)
14:46:50 <betehess> will be ..
14:47:18 <sandro> SteveS: If we can reuse syntax we should, community feedback is good, enumerate in draft the hot areas
14:47:32 <JohnArwe> q+
14:47:38 <Arnaud> ack pchampin
14:47:41 <sandro> SteveS: raise issues, and put them directly in WD
14:48:18 <sandro> pchampin: About the syntax - I also agree we should change the slice separator, because colon looked too much like pname, but dotdot would be a much better choice
14:48:24 <sandro> -1000 backslash
14:48:28 <betehess> ok ok
14:48:32 <betehess> :-)
14:48:47 <sandro> pchampin: lets do dotdot before fpwd
14:48:52 <SteveS> also, I believe we should try hard to reuse existing syntax/semantics but think we should highlight and get community feedback.  I haven’t gone through proposal in detail yet with my use cases
14:48:56 <Arnaud> ack JohnArwe
14:48:59 <betehess> so, let's publish with s/-/^/ and s/>/../
14:49:02 <ericP> backslashes suck in strings
14:49:29 <sandro> JohnArwe: charter extension is to 6 months
14:49:38 <betehess> ericP, they backsuck
14:49:45 <sandro> Arnaud: we're not thinking ld-patch will be done in this WG
14:50:02 <sandro> Arnaud: in re-chartered WG
14:51:00 <sandro> Arnaud: the hope is to recharter and continue it in followon WG
14:51:11 <sandro> Arnaud: It would be hard to get it to REC by the end of the year
14:51:23 <pchampin> q+
14:51:52 <sandro> Arnaud: Do we want to highlight in status that we're not 100% sure this is the way we want to go?
14:52:06 <Arnaud> ack pchampin
14:52:17 <sandro> +1 say that, yeah
14:52:34 <betehess> q+
14:52:44 <sandro> pchampin: Something that could help people understand the syntax.     Include a section on rewriting LD-PATCH into SPARQL update
14:52:46 <sandro> +1
14:52:54 <SteveS> +1, should highlight in status yes
14:53:12 <betehess> this is called denotational semantics, and I completely agree
14:53:19 <sandro> Arnaud: Yes, that'd be great
14:53:25 <SteveS> +1 to LD-PATCH->SPARQL update section
14:53:58 <sandro> ericP: If we wrote up sandro or my proposal, we could include a reference to them in the document
14:54:16 <JohnArwe> +1 to pchampin
14:54:31 <sandro> eric: If "SPARQL UPDATE" is unconstrainted, then people will like it more than it deserves
14:55:01 <betehess> so we are basically rewriting the constraints
14:55:13 <betehess> current text says: [[
14:55:14 <betehess> The LD Patch format described in this document should be seen as an "assembly language" for updating RDF Graphs. It is the intention to confine its expressive power to an RDF diff with minimal support for blank nodes. For more powerful operations on RDF Graphs and Quad Stores, the LDP WG recommends the reader to consider SPARQL Update.  ]]
14:55:44 <betehess> (should add support for list in this description as well)
14:55:54 <betehess> Arnaud, can we vote withing the next 5 minutes??
14:56:06 <betehess> publishing
14:56:10 <betehess> or not
14:56:15 <betehess> then move on
14:56:52 <sandro> PROPOSED: Publish LD-PATCH as FPWD with inverse path and slice syntax fixed, possibly other raised issues (eg slash syntax), links to sandro's and eric's proposal
14:56:55 <betehess> well, it's a draft, can be updated
14:57:37 <Ashok> Yes, that looks like a good option
14:57:47 <sandro> PROPOSED: Publish LD-PATCH as FPWD with inverse path and slice syntax fixed, possibly other raised issues (eg slash syntax), links to sandro's and eric's proposal (explaining we're asking feedback about which direction to go)
14:57:54 <TallTed> +1
14:57:59 <Ashok> +1
14:57:59 <SteveS> +1
14:58:00 <sandro> +0
14:58:00 <pchampin> +1
14:58:05 <betehess> +1
14:58:18 <JohnArwe> +1
14:58:25 <sandro> RESOLVED: Publish LD-PATCH as FPWD with inverse path and slice syntax fixed, possibly other raised issues (eg slash syntax), links to sandro's and eric's proposal (explaining we're asking feedback about which direction to go)
14:58:27 <ericP> +1
14:58:55 <betehess> ok, I will come up with an updated draft for tomorrow or Wednesday
14:59:25 <sandro> issue: should LD-PATCH use a slash like SPARQL does, instead of as it currently does?
14:59:26 <trackbot> Created ISSUE-100 - Should ld-patch use a slash like sparql does, instead of as it currently does?.  Please complete additional details at <http://www.w3.org/2012/ldp/track/issues/100/edit>.
14:59:30 <Zakim> -TallTed
14:59:31 <Zakim> -Alexandre
14:59:34 <Zakim> -Ashok_Malhotra
14:59:37 <Zakim> -Arnaud
14:59:47 <Zakim> -pchampin
14:59:49 <Zakim> -JohnArwe
14:59:51 <Zakim> -SteveS
14:59:55 <Zakim> -Sandro
14:59:55 <Zakim> -EricP.a
14:59:56 <Zakim> SW_LDP()10:00AM has ended
14:59:56 <Zakim> Attendees were pchampin, Arnaud, Ashok_Malhotra, SteveS, +1.857.928.aaaa, Alexandre, +1.857.204.aabb, ericP, Sandro, JohnArwe, TallTed
15:11:07 <bblfish> bblfish has joined #ldp
15:51:18 <bblfish> bblfish has joined #ldp
16:04:52 <SteveS> SteveS has joined #ldp
16:26:06 <Arnaud> zakim, make log public
16:26:06 <Zakim> I don't understand 'make log public', Arnaud
16:26:16 <Arnaud> zakim, make logs public
16:26:16 <Zakim> I don't understand 'make logs public', Arnaud
16:26:16 <bblfish> bblfish has joined #ldp
16:26:43 <Arnaud> RRSAgent, make logs public
18:36:24 <Zakim> Zakim has left #ldp
19:03:24 <bblfish> bblfish has joined #ldp
19:56:24 <deiu> deiu has joined #ldp
20:14:19 <deiu_> deiu_ has joined #ldp
21:56:53 <bblfish> bblfish has joined #ldp