13:58:16 <RRSAgent> RRSAgent has joined #ldp
13:58:16 <RRSAgent> logging to http://www.w3.org/2014/07/28-ldp-irc
13:58:18 <trackbot> RRSAgent, make logs public
13:58:18 <Zakim> Zakim has joined #ldp
13:58:20 <trackbot> Zakim, this will be LDP
13:58:20 <Zakim> ok, trackbot, I see SW_LDP()10:00AM already started
13:58:21 <trackbot> Meeting: Linked Data Platform (LDP) Working Group Teleconference
13:58:21 <trackbot> Date: 28 July 2014
13:58:25 <Zakim> +Sandro
13:59:11 <Zakim> +[IPcaller]
13:59:14 <Ashok> Ashok has joined #ldp
13:59:22 <betehess> Zakim, IPCaller is Alexandre
13:59:22 <Zakim> +Alexandre; got it
13:59:55 <Zakim> -Sandro
14:00:13 <Zakim> +Arnaud
14:00:24 <Zakim> +??P17
14:00:31 <deiu> Zakim, ??P17 is me
14:00:31 <Zakim> +deiu; got it
14:00:32 <Zakim> +Sandro
14:00:40 <deiu> Zakim, mute me please
14:00:40 <Zakim> deiu should now be muted
14:00:42 <Zakim> +JohnArwe
14:01:30 <Zakim> +Ashok_Malhotra
14:02:00 <Zakim> +OpenLink_Software
14:02:16 <TallTed> Zakim, OpenLink_Software is temporarily me
14:02:16 <Zakim> +TallTed; got it
14:02:18 <TallTed> Zakim, mute me
14:02:18 <Zakim> TallTed should now be muted
14:02:38 <deiu> Zakim, who is speaking?
14:02:49 <Zakim> deiu, listening for 11 seconds I heard sound from the following: Sandro (46%)
14:03:13 <sandro> better?
14:03:58 <Arnaud> zakim, who's on the phone/
14:03:58 <Zakim> I don't understand 'who's on the phone/', Arnaud
14:04:09 <Arnaud> zakim, who's on the phone?
14:04:09 <Zakim> On the phone I see Alexandre, Arnaud, deiu (muted), Sandro, JohnArwe, Ashok_Malhotra, TallTed (muted)
14:04:14 <nmihindu> nmihindu has joined #ldp
14:04:25 <JohnArwe> regrets: steve speicher, sergio
14:04:27 <Zakim> +Roger
14:04:38 <roger> roger has joined #ldp
14:04:49 <betehess> scribe: Alexandre
14:04:55 <betehess> scribenick: betehess
14:05:12 <betehess> Arnaud: let's start with approval of last minutes
14:05:23 <Zakim> +ericP
14:05:23 <betehess> ... any objection?
14:05:25 <deiu> minutes looked ok
14:05:50 <Arnaud> http://www.w3.org/2013/meeting/ldp/2014-07-21
14:05:51 <betehess> ... no objection, approved
14:06:08 <betehess> ... next meeting on Aug 4th
14:06:12 <betehess> ... no objection?
14:06:20 <betehess> ... ok, it works
14:06:28 <bblfish> bblfish has joined #ldp
14:06:32 <betehess> ... now actions
14:06:44 <JohnArwe> issue-99?
14:06:44 <trackbot> issue-99 -- Validate the JSON-LD examples of the primer -- closed
14:06:44 <trackbot> http://www.w3.org/2012/ldp/track/issues/99
14:07:06 <JohnArwe> email on 99 from nandana: http://lists.w3.org/Archives/Public/public-ldp-wg/2014Jul/0066.html
14:07:11 <betehess> ... let's close action-144
14:07:27 <betehess> JohnArwe: haven't heard back from issue-99
14:07:44 <betehess> ... sandro wanted to publish a json-ld context file in HG
14:07:55 <betehess> sandro: sorry, no idea what this is
14:07:56 <Zakim> +??P4
14:08:05 <JohnArwe> text of email: In Hydra spec examples, they use a JSON-LD context document,
14:08:05 <JohnArwe> http://www.w3.org/ns/hydra/context.jsonld , which correctly sets with the
14:08:05 <JohnArwe> CORS header "Access-Control-Allow-Origin: *" in the response. If we can do
14:08:05 <JohnArwe> the same, it solves the problem that the validators [1] have with our
14:08:06 <JohnArwe> current JSON-LD context file [2] in the Mercurial repository due to
14:08:06 <JohnArwe> cross origin issues. Can we publish a context file in a similar way?
14:08:16 <betehess> Arnaud: Greg Kellogs said that the json-ld examples were not correct
14:08:22 <betehess> ... we need to have a context
14:08:24 <nmihindu> Zakim, ??P4 is me
14:08:24 <Zakim> +nmihindu; got it
14:08:28 <betehess> sandro: I understand
14:08:39 <betehess> ... are we publishing context on w3.org?
14:08:43 <betehess> ... or dvcs.w3.org?
14:08:55 <betehess> ... may not be good for long term
14:09:08 <betehess> ... we can start with dvcs.w3.org for now
14:09:15 <JohnArwe> sounds like an action so we don't lose it?
14:09:34 <betehess> ... will raise the issue with json-ld folks
14:10:34 <betehess> Arnaud: has to be outside of the document if you don't want to repeat it
14:10:52 <betehess> ... sandro, can take the action to see w/ the json-ld folks?
14:10:56 <sandro> action: sandro figure out where json-ld context for LDP goes on w3.org
14:10:57 <trackbot> Created ACTION-145 - Figure out where json-ld context for ldp goes on w3.org [on Sandro Hawke - due 2014-08-04].
14:11:34 <betehess> ... anything else re: issues/actions?
14:11:51 <betehess> topic: Paging spec
14:12:05 <betehess> Arnaud: didn't have much time last week to discuss it
14:12:16 <betehess> ... paging was split of the main spec
14:12:23 <betehess> ... went through 1st LC
14:12:31 <betehess> ... want it to move forward
14:12:56 <betehess> ... has been discussion between sandro and JohnArwe
14:13:07 <betehess> ... haven't have a change to catch up with the ML, sorryu
14:13:13 <betehess> ... can I have a summary?
14:13:31 <betehess> JohnArwe: I have added links to the agenda
14:13:48 <JohnArwe> Steve Speicher had also provided a few comments 1 week + ago, which I agreed with in the large although I don't think I ever responded on the list saying so
14:14:39 <betehess> ... let's start going down the list
14:14:52 <betehess> .... "hijacking prefer"
14:15:05 <betehess> http://lists.w3.org/Archives/Public/public-ldp-comments/2014Jul/0000.html
14:15:19 <betehess> ... @@ is nervous about us hijacking Prefer
14:15:32 <betehess> ... different interpretation
14:15:48 <betehess> ... had a draft response in my email
14:16:07 <betehess> ... the Page size header is important
14:16:35 <betehess> ... sandro suggested to clarify that
14:16:38 <deiu> s/@@/Gregory McFall/
14:16:44 <betehess> ... it'd be clearer
14:17:17 <betehess> Arnaud: we should just clarify the spec
14:17:23 <betehess> ... and ask him if he is happy
14:17:33 <betehess> JohnArwe: no disagreement?
14:18:00 <betehess> Arnaud: people are comfortable w/ JohnArwe making edits to the spec to clarify the intent, and tell Gregory about that
14:18:22 <JohnArwe> action: johnarwe to clarify the spec, saying that our intent is that ldp paging support "signal" requires the presence of the 'pagsize' parameter, and to reply on the public comments list to see if that addresses the concern
14:18:22 <trackbot> Error finding 'johnarwe'. You can review and register nicknames at <http://www.w3.org/2012/ldp/track/users>.
14:18:40 <JohnArwe> action: john to clarify the spec, saying that our intent is that ldp paging support "signal" requires the presence of the 'pagsize' parameter, and to reply on the public comments list to see if that addresses the concern
14:18:40 <trackbot> Created ACTION-146 - Clarify the spec, saying that our intent is that ldp paging support "signal" requires the presence of the 'pagsize' parameter, and to reply on the public comments list to see if that addresses the concern [on John Arwe - due 2014-08-04].
14:18:58 <betehess> ... next is: Must Not > Should Not http://lists.w3.org/Archives/Public/public-ldp-comments/2014Jul/0002.html
14:19:21 <betehess> JohnArwe: it's about the wording on client signaling to the server
14:19:39 <betehess> ... initially MUST NOT, sandro proposed SHOULD NOT
14:19:46 <betehess> Arnaud: are we circling back?
14:20:12 <betehess> sandro: MUST NOT was because we want to avoid having broken clients everywhere
14:20:47 <betehess> ... but sometimes, the server knows it's a bad idea
14:20:47 <deiu> q+
14:20:55 <betehess> ... and would prefer to fail
14:21:03 <betehess> ... eg. the payload is too big
14:21:11 <betehess> ... hence SHOULD NTO
14:21:21 <ericP> i thought the issue was that we don't want the server to lie to the client
14:21:37 <sandro> its never lying eric
14:21:42 <betehess> Arnaud: that's similar to original argument
14:22:13 <ericP> it is if you ever care about the difference between the whole resource and a page of it
14:22:18 <betehess> ... this is not radical change
14:22:34 <TallTed> it's negotiation.... neither side is purely in control
14:22:42 <betehess> ... what to people think?
14:22:43 <ericP> i'd expect timbl to have issues with this too
14:22:46 <deiu> Is there a way then to explicitly tell the server that you want the whole thing?
14:22:59 <JohnArwe> personally, I'm fine with should.  I missed the original "must" add, and if you remember my reaction the next week, it was barely printable.
14:23:00 <TallTed> Zakim, unmute me
14:23:00 <Zakim> TallTed should no longer be muted
14:23:07 <deiu> Or is the server _always_ going to force paging because it's "smarter"
14:23:47 <deiu> q-
14:24:05 <betehess> sandro: the wording should not allow server to do paging if they think the client wants it
14:24:29 <TallTed> TallTed: HTTP spec wants the technology to figure things out on its own when it can, but has a "pass to user" when it can't, e.g., ConNeg
14:24:39 <betehess> ericP: I think timbl couldn't tell the difference between full resource and paged one
14:24:39 <JohnArwe> sandro: wants the sense to be "not whenever you want to page"
14:24:52 <TallTed> TallTed: This might be one of those places
14:24:58 <JohnArwe> we have 303 now
14:24:58 <betehess> ... how do we tell the client: here is the full thing, and a page of it
14:25:05 <JohnArwe> this is not about 2nn
14:25:11 <deiu> aha, ok
14:25:17 <betehess> sandro: this only applies to 303
14:25:23 <betehess> ... not 2nn
14:27:02 <JohnArwe> sandro: believe I was told that getting the value of 2nn could be done in 1 week outside of wg.
14:27:29 <ericP> Yves, ears on?
14:28:44 <betehess> Arnaud: back to MUST/SHOULD NOT
14:28:51 <TallTed> +1 SHOULD
14:28:55 <betehess> ... we have a proposal: moving back to SHOULD NOT
14:29:05 <JohnArwe> +1 should not
14:29:26 <ericP> action: ericP to follow up on Yves's way to turn 2NN into a number without waiting for the draft to go to RFC
14:29:26 <trackbot> Created ACTION-147 - Follow up on yves's way to turn 2nn into a number without waiting for the draft to go to rfc [on Eric Prud'hommeaux - due 2014-08-04].
14:29:30 <betehess> PROPOSAL: moving MUST NOT to SHOULD NOT
14:29:41 <betehess> +0
14:29:57 <deiu> +0 (as long as clients can request the full representation)
14:30:14 <sandro> PROPOSAL: Say that servers SHOULD NOT initiate paging by 303-redirecting to the rel=first page
14:30:24 <sandro> +1
14:30:36 <Ashok> +1
14:30:37 <betehess> +0
14:30:49 <sandro> PROPOSAL: Say that servers SHOULD NOT initiate paging by 303-redirecting to the rel=first page (in the case where the client DIDNT ask for it); this does NOT apply to 2nn
14:30:57 <betehess> ericP: you want to include that 2nn cannot be used there?
14:31:00 <deiu> +0
14:31:03 <JohnArwe> +1
14:31:03 <betehess> ... bah, should be fine
14:31:04 <sandro> +1
14:31:07 <nmihindu> +1
14:31:09 <TallTed> +1
14:31:10 <roger> +1
14:31:23 <sandro> RESOLVED: Say that servers SHOULD NOT initiate paging by 303-redirecting to the rel=first page (in the case where the client DIDNT ask for it); this does NOT apply to 2nn
14:31:47 <betehess> Arnaud: next is: Adding items/pages
14:32:01 <betehess> JohnArwe: part of it was about sorting in a container
14:32:48 <betehess> ... difference of interpretation between sandro  and I
14:33:06 <betehess> ... but we may be in phase, sandro would need to read my response
14:34:27 <betehess> sandro: a sorted container must add where it belongs, an unsorted container must add at the end
14:34:44 <sandro> sandro: sorted paging is always at the end, unsorted is always at the end
14:34:56 <sandro> sandro: sorted paging is always at the rigtht place, unsorted is always at the end
14:35:49 <betehess> sandro: there are some invariant wrt previous/next links
14:36:03 <betehess> ... it is sorted paging, not sorted container
14:37:02 <betehess> ... raises the question to rename from SortedContainer to SortedPaged
14:37:03 <sandro> sandro: what about using pageSortCriteria and pagingSortCriterion ?
14:37:17 <betehess> JohnArwe: things are randomly ordered on a given page
14:37:36 <betehess> Arnaud: that is not a property on the container
14:37:42 <betehess> ... (like that)
14:37:52 <betehess> JohnArwe: there would be not value on the container
14:38:41 <betehess> sandro: you only see the sorting when you are doing paging
14:38:52 <JohnArwe> LDP today *does not* guarantee that *representations* are sorted, and each page is a repn
14:39:09 <sandro>  pagingSortCriteria, pagingSortCriterion, pagingSortCollation
14:39:12 <betehess> ... there is no order in a specific graph
14:39:37 <sandro>  pagingSortPredicate, pagingSortOrder
14:39:42 <betehess> JohnArwe: on a single page, this is just RDF
14:41:02 <JohnArwe> which 'so much expressiveness' are you thinking of?  I think of paging as being very thin.
14:41:34 <betehess> Arnaud: we
14:41:47 <betehess> Arnaud: we're only talking about some renaming for now
14:42:03 <betehess> ... other discussions are interesting, but we're diverging
14:42:58 <deiu> Being able to specify a sorting criteria seems very useful, though it looks more like filtering to me.
14:43:03 <betehess> PROPOSAL: rename from SortedContainer to SortedPaged
14:43:30 <betehess> PROPOSAL: use pageSortCriteria and pagingSortCriterion
14:43:38 <betehess> s/PROPOSAL: rename from SortedContainer to SortedPaged//
14:44:06 <betehess> +0, with all my heart
14:44:06 <JohnArwe> +1 rename
14:44:23 <deiu> +0 (by the way, would clients expect ordered results after specifying a sort criteria?)
14:44:38 <sandro> PROPOSAL: rename the paging predicates and classes to be like pagingSortCritera, since it's about the paging, not about the container
14:44:40 <TallTed> +0.5
14:44:43 <sandro> +1
14:44:50 <betehess> +0
14:45:03 <sandro> RESOLVED: rename the paging predicates and classes to be like pagingSortCritera, since it's about the paging, not about the container
14:46:00 <betehess> sandro: the server says "btw, this is sorted that way"
14:46:11 <betehess> ... the client can say "btw, this is the sorting I like"
14:46:46 <Ashok> yes
14:46:47 <JohnArwe> sandro wants to use the SAME link syntax to allow the client to express its preference for sort criteria
14:47:41 <sandro> sandro: How about allowing the clients to request the server what sort order it wants?
14:48:01 <betehess> Arnaud: maybe this is a broader question
14:48:21 <betehess> ... that would be for sorting now, may not scale
14:48:28 <betehess> +1 to Arnaud
14:49:18 <betehess> Arnaud: we can document it for next version
14:50:45 <betehess> Arnaud: anything else on paging to discuss now?
14:53:36 <sandro> JohnArwe: if we include the link request header, we might get some pushback
14:53:54 <betehess> Arnaud: we have time constraints
14:53:57 <sandro> Arnaud: I'm going to rule out-of-scope the client requesting the sort order
14:54:15 <betehess> ... what does it take to go for LC?
14:54:30 <betehess> JohnArwe: steve and sandro have been providing feedback
14:55:05 <betehess> sandro: we still have 2nn at risk
14:55:16 <betehess> ... people can't really implement today
14:56:09 <betehess> Arnaud: a few changes have to be made to the spec
14:56:12 <sandro> sandro: it would be nice to settle on 209 instead of 2nn so people can try implementing it!
14:56:18 <betehess> ... JohnArwe will tell us when it's done
14:56:39 <betehess> ... then we can decide to go LC last week
14:56:48 <betehess> ... that's last chance for people to review
14:57:21 <betehess> Arnaud: wanted to talk on Access Control spec
14:57:33 <JohnArwe> topic: access control spec
14:57:44 <betehess> Ashok: one feedback was we have no fine grained access control
14:58:09 <betehess> ... another comment was about resources
14:58:42 <betehess> Arnaud: first question is interesting, as it's about the meaning of publishing this
14:58:59 <betehess> ... the NOTE is about what we'd want to see in access control
14:59:03 <betehess> ... it's not binding us
14:59:20 <betehess> ... next group would be free to ignore it, it's just an input
15:00:24 <betehess> ericP: coming from outside, fine grained would mean dealing with triples
15:00:39 <betehess> Arnaud: are we ready to publish this document?
15:00:53 <betehess> ... I guess you'd prefer to hear from henry
15:01:09 <betehess> sandro: how are we speaking about fine grain?
15:01:15 <betehess> Arnaud: was just a comment
15:01:46 <betehess> ... it depends on definition of fine grain
15:02:01 <betehess> sandro: you have a url on everything you are setting permission on
15:02:27 <betehess> ... this is confusing because URL can be anything
15:02:42 <ericP> if i go to a linkedin page and i'm logged in, i see more than if i'm not
15:02:46 <betehess> Arnaud: we're out of time
15:02:54 <betehess> ... we can get an updated version
15:02:56 <ericP> i'd characterize that as fine-grained
15:03:23 <betehess> Arnaud: people have been talking on ML about LD Patch format
15:03:28 <betehess> ... crazy people
15:03:39 <betehess> ... we'll get back to Patch next
15:03:50 <Zakim> -Ashok_Malhotra
15:03:54 <betehess> ... and we'll try to publish Paging as well
15:03:55 <TallTed> fine-grained is in the eye of the beholder/implementer.  for purposes of ACLs, fine-grained is the smallest thing you can identify, which in this world is the smallest thing you can specify by URI.
15:03:56 <Zakim> -ericP
15:04:00 <Zakim> -nmihindu
15:04:02 <Zakim> -Sandro
15:04:05 <Zakim> -deiu
15:04:06 <Zakim> -TallTed
15:04:06 <Zakim> -JohnArwe
15:04:07 <Zakim> -Roger
15:04:07 <Zakim> -Alexandre
15:04:09 <Zakim> -Arnaud
15:04:09 <Zakim> SW_LDP()10:00AM has ended
15:04:09 <Zakim> Attendees were Sandro, Alexandre, Arnaud, deiu, JohnArwe, Ashok_Malhotra, TallTed, Roger, ericP, nmihindu
15:04:36 <ericP> TallTed, so i think you've used your definition to say that linked-in acls are out of scope
15:05:09 <ericP> (i also thinkg that most folks would interpret document-level acls as course-grained)
15:05:10 <TallTed> the variance in what you see on a LinkedIn page isn't the *page* but the *elements on that page*
15:05:38 <ericP> if it were RDF, the variance would be that logged-in, you see more triples
15:06:46 <TallTed> yes, you see more triples -- but the ACLs are on the resources which comprise the composite resource (container) you're GETting
15:07:46 <TallTed> per-triple ACLs are kind of like saying "you can have the odd numbered slices of bread in this loaf"
15:08:36 <TallTed> or, "you can't have slices #5, #7, or #12, but here's the rest of the loaf"
15:29:09 <deiu> deiu has joined #ldp
15:58:24 <jmvanel> jmvanel has joined #ldp
16:28:29 <Zakim> Zakim has left #ldp