14:57:30 <RRSAgent> RRSAgent has joined #ldp
14:57:30 <RRSAgent> logging to http://www.w3.org/2013/12/09-ldp-irc
14:57:32 <trackbot> RRSAgent, make logs public
14:57:32 <Zakim> Zakim has joined #ldp
14:57:34 <trackbot> Zakim, this will be LDP
14:57:35 <trackbot> Meeting: Linked Data Platform (LDP) Working Group Teleconference
14:57:35 <trackbot> Date: 09 December 2013
14:57:36 <Zakim> ok, trackbot; I see SW_LDP()10:00AM scheduled to start in 3 minutes
14:57:53 <pchampin> pchampin has joined #ldp
14:58:30 <Zakim> SW_LDP()10:00AM has now started
14:58:37 <Zakim> +??P4
14:58:44 <pchampin> zakim, ??P4 is me
14:58:44 <Zakim> +pchampin; got it
14:59:36 <stevebattle8> stevebattle8 has joined #ldp
15:00:16 <Zakim> +Arnaud
15:00:31 <JohnArwe> JohnArwe has joined #ldp
15:00:41 <codyburleson> codyburleson has joined #ldp
15:00:42 <Zakim> + +1.845.454.aaaa
15:00:50 <JohnArwe> zakim, aaaa s me
15:00:50 <Zakim> I don't understand 'aaaa s me', JohnArwe
15:00:55 <Zakim> +SteveS
15:01:21 <JohnArwe> zakim, aaaa is me
15:01:21 <Zakim> +JohnArwe; got it
15:01:24 <Zakim> +[IPcaller]
15:01:33 <codyburleson> Zakim, IPCaller is me
15:01:33 <Zakim> +codyburleson; got it
15:01:35 <Zakim> +Sandro
15:02:39 <Zakim> +Alexandre
15:02:41 <Ashok> Ashok has joined #ldp
15:02:42 <codyburleson> Hear no audio. Is anyone talking?
15:02:50 <codyburleson> There it is.
15:03:05 <sandro> sandro has joined #ldp
15:03:09 <Zakim> +SteveBattle
15:03:18 <SteveS> I can only do 1st 60 minutes, scheduling conflict
15:03:26 <Zakim> +Ashok_Malhotra
15:05:03 <Arnaud> zakim, who's on the phone?
15:05:03 <Zakim> On the phone I see pchampin, Arnaud, JohnArwe, SteveS, codyburleson, Sandro, Alexandre, SteveBattle, Ashok_Malhotra
15:06:56 <roger> roger has joined #ldp
15:07:49 <Zakim> +bblfish
15:08:08 <Zakim> +OpenLink_Software
15:08:12 <Zakim> +Roger
15:08:14 <bblfish> bblfish has joined #ldp
15:08:17 <bblfish> hi
15:08:22 <TallTed> Zakim, OpenLink_Software is temporarily me
15:08:22 <Zakim> +TallTed; got it
15:08:26 <TallTed> Zakim, mute me
15:08:26 <Zakim> TallTed should now be muted
15:08:30 <betehess> Zakim, mute bblfish
15:08:30 <Zakim> bblfish should now be muted
15:08:47 <sandro> scribe: sandro
15:08:52 <bblfish> zakim, unmute bblfish
15:08:52 <Zakim> bblfish should no longer be muted
15:09:14 <sandro> Arnaud: Welcom everybody.   I hope we're close to resolving the few remaining issues.
15:09:20 <sandro> topic: admin
15:09:34 <roger> zakim, mute me
15:09:34 <Zakim> Roger should now be muted
15:09:42 <betehess> they look fine to me
15:10:17 <sandro> PROPOSED: approve minutes https://www.w3.org/2013/meeting/ldp/2013-12-02
15:11:23 <sandro> sandro: clarify closing issue-85?
15:11:45 <sandro> arnaud: I'll amend to clarify nothing needed to be done.
15:11:46 <sandro> +1
15:11:49 <betehess> no objection
15:11:55 <sandro> RESOLVED: approve minutes https://www.w3.org/2013/meeting/ldp/2013-12-02
15:12:06 <sandro> RESOLVED: approve minutes https://www.w3.org/2013/meeting/ldp/2013-12-02 (with minor amendment)
15:12:12 <sandro> Arnaud: Meeting next week, dec 16
15:12:21 <bblfish> q+
15:12:30 <sandro> .. we'll discuss then whether to meet on Dec 23
15:12:32 <Arnaud> ack bblfish
15:12:49 <sandro> bblfish: Regarding next F2F?     I can try to organize something in Paris
15:13:31 <roger> +1
15:13:40 <sandro> Arnaud: The bigger question is when....       Depends on LC comment period
15:13:59 <JohnArwe> since feb is now poss, I cannot do February 23-26, 2014
15:14:01 <Ashok> Can we have f2f in March?
15:14:11 <sandro> Arnaud: Cody also offered to host in Mexico.    When we figure out the time, we'll know better.
15:14:20 <codyburleson> Monterrey, Mexico
15:14:23 <bblfish> Ok. I can ask around and let you know what options exist for Paris
15:15:04 <sandro> bblfish, I've had F2F meetings at the ILOG building, which is presumably now an IBM building.
15:15:11 <sandro> subtopic: actions
15:15:21 <bblfish> also have connections at Mozilla Paris
15:15:23 <sandro> Ashok: I got some email about access control
15:15:57 <sandro> Ashok: Miguel is returning now, and suggests trying use cases
15:16:16 <bblfish> Yes, I'll try to add something on Access Control Use Cases too. (I have been wanting to do that for a while)
15:17:07 <sandro> Arnaud: 3 main issues remain
15:17:19 <sandro> Arnaud: We have three types of containers now.
15:17:28 <bblfish> Issue-90?
15:17:28 <trackbot> Issue-90 -- An LDPC/LDPR is a Named Graph -- open
15:17:28 <trackbot> http://www.w3.org/2012/ldp/track/issues/90
15:17:32 <bblfish> http://www.w3.org/2012/ldp/wiki/Issue-90
15:17:37 <sandro> topic: ISSUE-90 An LDPC/LDPR is a Named Graph
15:18:15 <sandro> Arnaud: Let's try to keep the issue separate, and I think it makes most sense to do ISSUE-90 first.
15:18:31 <sandro> Arnaud: Not we have the Graph Store Protocol, that already leverages Named Graphs
15:18:47 <sandro> .. so there is a precedent
15:19:01 <sandro> .. and we've already agreed we should be compatible when that works
15:19:09 <TallTed> s/Not we have/Note we have/
15:19:50 <betehess> http://www.w3.org/2012/ldp/wiki/Issue-90
15:20:01 <sandro> betehess: I hope everyone had time to read http://www.w3.org/2012/ldp/wiki/Issue-90
15:20:11 <sandro> betehess: Here are the two proposals to address this issue.
15:20:13 <JohnArwe> link's not in the issue either
15:20:58 <sandro> betehess: Main thing is LDPG is a more specifc LDPR that's a graph.
15:21:18 <sandro> .. non-LDPR is called an LDPB
15:21:18 <Kalpa> Kalpa has joined #ldp
15:21:42 <sandro> .. When you interact with an LDPC, you are always creating an LDPR, it's just one of these three.
15:21:59 <SteveS> q+
15:22:11 <sandro> Ashok: What would be the difference between a "container" and a "named graph"
15:22:15 <sandro> betehess: I'll come to that.
15:22:19 <Arnaud> ack steves
15:22:32 <sandro> SteveS: In your new defn of LDPR, does that include all resources on the web?
15:22:34 <JohnArwe> q+
15:22:41 <sandro> betehess: No.
15:23:00 <sandro> betehess: LDPRs are contained in an LDPC.
15:23:16 <sandro> Arnaud: This is a type of resource?
15:23:35 <sandro> betehess: Yes, created by interacting with container.
15:23:40 <codyburleson> +q
15:23:41 <sandro> q+
15:23:48 <Arnaud> ack JohnArwe
15:24:10 <sandro> JohnArwe: You said when you post to a container you create an LDPR
15:24:21 <sandro> .. what about a binary, with a metadata graph on the side
15:24:40 <sandro> betehess: for ISSUE-89 I said the LDPR is always the same, and the metadata is linked, it's not in the container.
15:24:55 <sandro> JohnArwe: Agreed, that's what's in the spec.
15:25:10 <sandro> betehess: The metadata is not an LDPR.
15:25:30 <sandro> betehess: Yes, this changes the defn from the spec of what an LDPR is.
15:25:56 <sandro> SteveS: The current spec an LDPR doesn't have to be in a container
15:26:01 <sandro> betehess: RIght.
15:26:05 <Arnaud> ack codyburleson
15:26:06 <bblfish> q+
15:26:50 <sandro> codyburleson: I was wondering this the other days.     "A successful post... results in a Named Graph".      Are you saying the container turns into a name graph?
15:26:59 <JohnArwe> q+ to ask about non-post creates of various sorts
15:27:03 <sandro> codyburleson: Can we make the container be a named graph?
15:27:13 <sandro> betehess: THat's part of proposal two.
15:27:20 <sandro> Arnaud: Let's work on the hierarchy first
15:28:12 <Arnaud> ack
15:28:15 <Arnaud> q?
15:28:31 <Arnaud> ack sandro
15:29:14 <sandro> sandro: what about non-contained LDPC?
15:29:39 <sandro> betehess: Ah, yes, this is an error.
15:29:50 <Arnaud> ack bblfish
15:29:54 <codyburleson> I have a proposal/idea for a "default root" container strategy, by the way; I think we need to address this - if not in spec - at least a "recommended approach". This is needed for pure LDP implementations.
15:29:59 <bhyland> bhyland has joined #ldp
15:30:00 <sandro> betehess: My focus is on what's created when you post to an LDPC
15:30:27 <betehess> created resource *implies* it is an LDPR
15:30:37 <sandro> bblfish: (not scribed -- couldnt follow)
15:31:11 <sandro> Arnaud: This is trying to label the binary ones.
15:31:17 <sandro> q+
15:31:22 <sandro> q-
15:31:26 <Arnaud> ack JohnArwe
15:31:26 <Zakim> JohnArwe, you wanted to ask about non-post creates of various sorts
15:31:33 <bblfish> agree this is just a matter to try to name some containers
15:31:42 <bblfish> s/containers/types of resources/
15:31:46 <sandro> JohnArwe: Is that true?    That this is mostly about naming LDPBs?
15:31:47 <sandro> q+
15:31:56 <sandro> betehess: Yes
15:32:08 <sandro> JohnArwe: Does this go beyond POST, to non-http methods, out of band?
15:32:28 <sandro> JohnArwe: I can have a container that is maintained out-of-band, it's a view over some rdbms table
15:32:44 <sandro> JohnArwe: When something is added by that out of band process, does this all apply
15:33:11 <bblfish> q+
15:33:18 <sandro> betehess: do you have an example.     it's hard to talk about things not created through LDP spec.     I don't know
15:33:51 <sandro> JohnArwe: A list of all my bugs, in a database, viewed via LDPC.
15:34:10 <sandro> betehess: Do you make a distinction between membership and containment?
15:34:13 <JohnArwe> s/JohnArwe:/SteveSpeicher:/
15:34:41 <sandro> bblfish: The issue here is to distinguish certain types of resources.
15:35:11 <SteveS> q+
15:35:12 <Arnaud> ack sandro
15:35:13 <sandro> Arnaud: This is about more accurate naming in the spec.
15:35:14 <pchampin> q+
15:35:56 <betehess> I'm fine with any isomorphic mapping
15:36:03 <betehess> con't care about names
15:36:09 <betehess> s/con/don/
15:36:48 <bblfish> q?
15:36:49 <sandro> sandro: "LDPG" is LDPR-that-is-RDF-Source,   "LDPB" is LDPR-that-is-not-RDF-source
15:36:49 <bblfish> q-
15:36:58 <pchampin> q-
15:37:11 <JohnArwe> thx for starting from "familiar" terms Alexandre, even if they're not final tnhey're a useful bridge
15:37:15 <Arnaud> ack steves
15:37:47 <sandro> SteveS: To be more accurate in the defn, an LDPR is an LDPG-or-LDPB, and the LDPC is an LDPG.
15:38:05 <sandro> bete: Maybe, yeah.
15:38:51 <sandro> Arnaud: It's not clear how much of LDPR an LDPC inherits.
15:39:46 <sandro> SteveS: So maybe LDPR and LDPC are disjoint
15:40:04 <sandro> sandro: No, they're just partially overlapping, so LDPC can be in LDPC
15:40:05 <JohnArwe> "not" disjoint
15:40:31 <sandro> Arnaud: Hopefully we've discussed this enough that people understand what we're trying to achieve.
15:40:46 <bblfish> btw, I had made a proposal similar to this some time ago. http://www.w3.org/2012/ldp/wiki/images/b/b5/LDP-RCX.pdf
15:40:56 <sandro> topic: ISSUE-90, Proposal-2 "LDP Named Graphs"
15:41:12 <sandro> betehess: This is about where the triples go when you add them, eg via POST
15:41:46 <sandro> TallTed, yes, a venn diag would be good
15:42:00 <Ashok> +1
15:42:27 <sandro> betehess: The LDPC would have some information about containment, as in ISSUE-89.   What's not clear in spec -- membership triples belong to COntainer Resource.
15:42:46 <sandro> .. This spec says where the triples are supposed to be.
15:42:53 <sandro> s/spec/proposal/
15:43:02 <JohnArwe> q+
15:43:07 <Arnaud> ack JohnArwe
15:43:31 <SteveS> q+
15:43:54 <sandro> JohnArwe: I think you;'re contraining how the membership and containment triples are stored, what named graph they're part of.    that seems like it might be overreaching.
15:44:27 <sandro> JohnArwe: As an example, David Wood has said Callimachus puts all the server-managed properties in the same named graph as the application data.
15:44:32 <Arnaud> ack steves
15:44:52 <sandro> .. It sounds like you're saying there must be a single NG that contains all the containment and membership properties.   THat might not be compatible.
15:44:54 <Arnaud> q=steves
15:45:03 <Arnaud> queue=steves
15:45:04 <sandro> .. it seems wrong to constrain where things are stored.
15:45:29 <sandro> betehess: This is just about the server managed triples.  It doesn't forbid application-managed triples in the same graph.   The server just has to know which is which.
15:46:09 <Ashok> q+
15:46:12 <sandro> betehess: And: Can the triples exist in multiple locations?  Yes, but that's outside this proposal.   THere's not expectation, after a DELETE, that a specific triple would disappear if it happens to be somewhere else.
15:46:36 <sandro> s/THere's not/There's no/
15:47:01 <sandro> SteveS: If you're not attempting to constrain where the triples are served, then perhaps it's okay.
15:47:25 <sandro> .. THe question is what is the Bare Minimum.     The LDPC MUST contain the membership and containment triples.
15:47:31 <SteveS> s/SteveS/JohnArwe/
15:47:42 <sandro> betehess: this is about interface, not storage
15:48:05 <Arnaud> ack steves
15:48:10 <bblfish> q?
15:48:33 <sandro> SteveS: The member says what container it's part of.
15:49:10 <sandro> SteveS: You're saying there's an NG, the Container Resource, and you're saying it contains...    the In-Scheme relation ... is in That Graph?
15:49:10 <sandro> betehess: Yes.
15:49:33 <sandro> SteveS: So in your example 5 if you used TriG, then it could be enforced
15:49:52 <JohnArwe> cygri's ex: member resource includes triple T = <member, skos:inScheme, scheme-uri>
15:49:54 <sandro> betehess: I've rewritten example 5 with several curl requests
15:50:08 <sandro> .. that's equiv to TriG
15:50:16 <sandro> SteveS: No, it's not.    That's not testable.
15:50:46 <sandro> betehess: The spec doesn't say anythign about where the triples are.   The member triples could still be part of the LDPG.
15:51:26 <sandro> Arnaud: In Ex 5 do you still get the liability and asset containers?
15:51:56 <Arnaud> https://dvcs.w3.org/hg/ldpwg/raw-file/default/ldp.html#ldpc-ex-membership-full
15:52:04 <sandro> SteveS: Ex 5 shows the first curl...      How did the client know to get that URI?
15:52:25 <bblfish> q+
15:52:27 <sandro> SteveS: If you add example 5 B, curl get on TriG, THEN you can highlight the NG and differences.
15:52:31 <JohnArwe> in spec ex 5, single GET returns netWorth asset + asset container triples + liab triples (shown on wiki page as 3 GETs)
15:53:06 <bblfish> q-
15:53:11 <sandro> betehess: I don't see any problem when you do a get on the container resource ...    the spec just says okay, but they don't belong to the container resource, BUT IT DOESNT MEAN YOU CANT RETURN THEM AS WELL.
15:53:25 <bblfish> q+
15:53:28 <sandro> betehess: I'm fine with using TriG to make that explicit in the example.
15:53:45 <sandro> Arnaud: This touches on the inlining question.   inlining would be well-addressed by using TriG
15:54:26 <Arnaud> ack ashok
15:54:38 <sandro> Arnaud: GET returns a representation of NetWorth, if the server also wants to send along stuff about the asset container, that's okay.   I don't think I want the spec relying on TriG, even for documentation purposes.
15:55:09 <sandro> Ashok: I'm still trying to figure out what the relation is between a container and a graph.   I think you said a named graph can have a container and other LDPRs?
15:55:28 <JohnArwe> if worried about appearance of trig, could just show it both ways: as alexandre's wiki page has, and as spec currently has, both legal
15:56:20 <sandro> betehess: The named graph is just a URI and an RDF Graph.   It's important...   when I read Arnaud's example with SPARQL, it didn't make sense since it assumed everything was part of the same Named Graph.    I'd think when I post something, I end up with another graph with another name.   When I have to remove a membership triple, I need to know where to look,.
15:56:29 <SteveS> I agree with JohnArwe, or we could even use JSON-LD ;)
15:56:49 <sandro> Ashok: Can a named graph have one or more containers or other resources?
15:57:07 <sandro> Ashok: Can a named graph have a container AND other resources?
15:57:11 <SteveS> I struggle with how we write this normatively, using named graphs this way sounds like good implementation guidance / best practice
15:57:15 <sandro> .. that are not within the container.
15:57:33 <sandro> betehess: What does it mean for a container to have a named graph?
15:58:03 <sandro> betehess: A named graph is just a pair of URI and graph.   Maybe you mean a Graph Store (a collection of named graphs).
15:58:07 <sandro> q?
15:58:10 <Arnaud> ack bblfish
15:58:24 <bblfish> https://gist.github.com/anonymous/66c779451c94f90d271e
15:58:27 <sandro> bblfish: I think one can remove some of the issues here.   I just posed a gist.
15:58:57 <sandro> bblfish: People are asking if you can put information in membership triple graph about the containers.
15:59:17 <sandro> bblfish: You don't need TriG for this.
15:59:39 <sandro> bblfish: I've added a description of the asset container and the liability container.
16:00:07 <Zakim> -SteveS
16:00:36 <sandro> Arnaud: To me, your gist looks exactly like what's in the spec.   Are you just telling us you can still do that?
16:00:41 <JohnArwe> I think Ashok is asking if a Single named graph's representation "be/have" both a container and other resources... which is hard to answer b/c we might look at that question from a data standpoint only, from an interaction  model view only, or both.
16:01:00 <sandro> bblfish: There's some info about containers.     No need for trig.
16:01:17 <betehess> a good practice would be to have a triples from the ContainerResource to the LDPCs, living in the ContainerResource, so that the LDPCs are discoverable
16:01:21 <sandro> Arnaud: Is what your gist still legal in betehess's new world?
16:01:51 <JohnArwe> I've seen words in recent minutes appearing to say that a resource's type determines its interaction model, which strike me as "there be dragons"
16:02:23 <sandro> betehess: If you want to follow GSP, that would be illegal
16:02:42 <sandro> betehess: But I don't say anything about that.
16:02:53 <JohnArwe> good catch sandro, I *was* hearing "legal"
16:03:39 <sandro> Arnaud: RIght now the spec allows all this, but you might be containing to stop it.
16:04:21 <JohnArwe> q+
16:04:31 <sandro> betehess: I think it should be allowed, but a bad practice.
16:04:49 <Arnaud> ack JohnArwe
16:04:57 <sandro> Arnaud: Others will call it good practice, saving the client several round trips.
16:05:49 <sandro> JohnArwe: It should be possible, if it is useful to us, to write some of our language in terms of named graphs, without bringing in the whole GSP.    I have the impression that's where Alexandre is going.
16:06:21 <sandro> betehess: Yes.    It doesn't define the GET.    This is only about the Named Graph.
16:06:26 <sandro> (I have no idea what that means)
16:06:47 <JohnArwe> so what's the next step  on 90?
16:06:52 <betehess> sandro, it means that the GET may return *more* that the Graph part of the Named Graph
16:07:21 <sandro> Arnaud: Next step on ISSUE-90 is let people think about it.
16:07:52 <bblfish> Issue-89?
16:07:52 <trackbot> Issue-89 -- Tie the interaction model with the LDP data model through the notion of Managed Resources -- open
16:07:52 <trackbot> http://www.w3.org/2012/ldp/track/issues/89
16:08:00 <sandro> JohnArwe: Maybe betehess wants to update his proposals based on discussion so far.
16:08:07 <Arnaud> http://www.w3.org/2012/ldp/wiki/Issue-89
16:08:09 <bblfish> http://www.w3.org/2012/ldp/wiki/Issue-89
16:08:15 <sandro> topic: ISSUE-89
16:08:35 <sandro> betehess: Call these Proposals 3, 4, 5
16:09:48 <sandro> .. So this just adds new triples, ldp:contains, and in the case of the simple container it would amend the proposal from two weeks ago so the simple container, the membership and containment triples are the same.
16:10:24 <sandro> betehess: Lata para, LDPC has RDF representation, with Link hear.
16:10:44 <JohnArwe> q+ to ask about indirect case
16:10:49 <bblfish> I renamed the proposals
16:10:53 <bblfish> to 1,2,3
16:11:09 <sandro> sandro: You mean Link Header, rel=type, NOT rel-Link
16:11:18 <sandro> betehess: I'll fix that.
16:11:23 <JohnArwe> henry, alexandre had said he wanted 3-5
16:12:25 <sandro> Arnaud: I the case of the simple container, this just renames the predicate.   It puts to rest what ought to be called "membership", and saying there's membership AND containment.
16:14:00 <sandro> Arnaud: In the case of the Indirect Container, there's two relation.  In the case of Direct Container, we only have one relation.     Direct Container to Member Resource.  You can INFER this Contains relationship.  Is it okay to keep it as something you can infer?   Earlier we said yes, but now betehess is proposing there be two triples.
16:14:16 <JohnArwe> q?
16:14:54 <Arnaud> ack john
16:14:54 <Zakim> JohnArwe, you wanted to ask about indirect case
16:14:55 <sandro> Arnaud: Porposal 2 is about the same as ldp:created.  Two things.   We've been talking about two relations all along.   Sometimes they happen to be the same.   This is materializing both.
16:15:13 <sandro> JohnArwe: On proposal 1, I want to understand.  In the indirect case, how does this work?
16:15:37 <sandro> betehess: The location header is about the created resource, not the object of the membership triple.
16:16:35 <JohnArwe> fwiw glad to see we agreed that this distinction betw membership and created/contains has crystallized
16:16:44 <sandro> betehess: Proposal 3 -- this needs more work.
16:16:53 <sandro> .. it depends on gathering some use cases.
16:17:39 <sandro> .. if you add ldp:contains triples, in some cases you double the number of triples in the LDPC.    The worst case, a direct container, the membership subject is still the ldpc.
16:18:04 <sandro> .. so trying to deal with that.   non-member resource.
16:18:14 <sandro> .. in this proposal, you just ask for the properties of the resource.
16:18:24 <sandro> .. for this one, I invite people to think about this a little more.
16:18:29 <sandro> q?
16:18:50 <sandro> .. this proposal is in the spirit of the non-member resource approach
16:19:11 <JohnArwe> q+ to talk about use case
16:19:13 <sandro> Arnaud: We don;t need to get into the gory details.  Once we materialize the containment relation, in some cases we end up with double the triples.
16:19:51 <sandro> .. so this proposal is a revamp of non-member-property, that says we're going to give you a filtering mecahnism to not have to see the duplication.
16:19:53 <sandro> q+
16:20:11 <Arnaud> ack john
16:20:11 <Zakim> JohnArwe, you wanted to talk about use case
16:20:27 <sandro> JohnArwe: Conceptually, I'm okay with this, if we can truly get rid of duplicates
16:21:00 <betehess> in a nutshell: a GET on a DirectContainer could return both membership triples *and* containment triples. The idea would be to either filter out or ask explicitly some of the triples eg. property triples
16:21:01 <sandro> JohnArwe:We have implementations that plan to use this stuff, and the problem today is scale & performance.    We've gotten up to millions of members.   And we have throughput requirements.
16:21:20 <Arnaud> ack sandro
16:21:23 <sandro> JohnArwe: So it's important to avoid the duplication.  Can't afford the 50% hit.
16:21:49 <sandro> sandro: Stupid question -- why do I want membership triples if I have containment triples.
16:21:49 <TallTed> Zakim, unmute me
16:21:49 <Zakim> TallTed should no longer be muted
16:21:49 <bblfish> q+
16:23:08 <SteveS> I think for DirectContainers it seems like too much overhead to ask to generate the duplication, if clients of direct containers need this..then they can generate it before handing off to some generic processing
16:23:11 <TallTed> containment implies membership.  membership does not imply containment.
16:23:24 <sandro> Arnaud: You either force the app to change its data, or you can say the direct container type of thing says how to client should interpret the data.
16:23:35 <JohnArwe> Membership triples are ALL we normally care about.  And most of them would not be  not created through containers.
16:23:39 <TallTed> this does require some reasoning
16:24:37 <sandro> Arnaud: Look at the three different types of containers
16:24:44 <sandro> q?
16:24:46 <TallTed> Zakim, mute me
16:24:46 <Zakim> TallTed should now be muted
16:24:49 <JohnArwe> ...that's why i was probing on out of band creates/deletes before.
16:25:17 <Arnaud> ack bblfish
16:25:25 <TallTed> wishing more for graphical representations.  the intended meanings of the words we use are often very difficult to interpret.
16:25:45 <sandro> bblfish: You put the ldp:contains in the ldpc, then you have another resource that contains the membership triples.
16:25:52 <sandro> bblfish: works for me.
16:25:59 <sandro> Arnaud: Needs more examples.
16:26:05 <bblfish> and a link to the membership triples examples
16:26:13 <bblfish> s/examples/resource/
16:26:31 <sandro> Arnaud: betehess please improve the wiki pages and email when done.
16:26:51 <sandro> betehess: I need more input on issue-89-proposal-3 before I do more on it.
16:26:59 <sandro> Arnaud: Show how it will play out.
16:27:23 <bblfish> q+
16:27:32 <sandro> sandro: Can;t you just use owl:samePropertyAs ?
16:28:09 <bblfish> http://www.w3.org/2012/ldp/wiki/Membership#Determining_the_membership_triples_to_be_added_when_a_new_member_is_created
16:28:13 <sandro> sandro: That's a standard way to ask for this kind of client-side processing
16:29:11 <sandro> bblfish: Even though it sounds like inferencing, sandro, from ldp:contains to some other predicate --- there's another way of interpreting it, where these things are only about creation.
16:29:41 <JohnArwe> I didn't quite say unaccept to materialize them Sandro; if we did something like the non-memb props trick (define new resource), so that our clients just never ask for them, might be fine.
16:29:45 <bblfish> here: http://www.w3.org/2012/ldp/wiki/Membership#Determining_the_membership_triples_to_be_added_when_a_new_member_is_created
16:29:58 <sandro> Arnaud: I added the SPARQL CONSTRUCT example to show how to do this.
16:30:22 <sandro> Arnaud: If we could have F2F *now* it would help us get through all this.
16:30:28 <bblfish> I'll look for Paris meeting
16:30:41 <sandro> sandro: We could try a remote all-day meeting.
16:30:51 <sandro> Arnaud: Maybe next week we can do 2-hours.
16:30:59 <JohnArwe> if we can get to agreement before people poof for year, I can draft specs while they're gone
16:31:26 <JohnArwe> being sick/injured now, i'll be here thru holidays mostly
16:31:51 <bblfish> I think we are making good progress
16:31:58 <sandro> Arnaud: I'm really trying to get us to closure here.  We're getting behind schedule.      This could be a problem soon.
16:32:18 <sandro> Arnaud: People need to make the effort to understand these issues.
16:32:23 <bblfish> very good progress. At least now we are starting to have vocabulary to speak of the issues, the SPARQL constructs are very helpful.
16:32:59 <sandro> Arnaud: If we don't make a decision to have a F2F soon, then we're deciding not to have it soon.
16:33:09 <betehess> if the process says "meeting", we could have a "gathering" instead :-)
16:33:38 <sandro> Arnaud: If we're luck we'll finish the issues this month, have LC for February, so F2F would be around March 1.  Is that okay?
16:33:50 <sandro> Arnaud: Are people okay for that timeframe?
16:33:52 <JohnArwe> once again: I will not make any f2f feb 23-26
16:34:14 <Ashok> How about early March?
16:34:18 <Arnaud> q?
16:34:21 <bblfish> q-
16:34:24 <sandro> Week of March 17?
16:34:31 <JohnArwe> I'm fine few weeks on either side of that
16:34:37 <sandro> ADJOURN
16:34:43 <Zakim> -TallTed
16:34:43 <Zakim> -bblfish
16:34:45 <betehess> bye!
16:34:45 <Zakim> -Ashok_Malhotra
16:34:47 <Zakim> -Alexandre
16:34:47 <Zakim> -Roger
16:34:48 <Zakim> -Sandro
16:34:49 <Zakim> -codyburleson
16:34:51 <Zakim> -Arnaud
16:34:53 <Zakim> -JohnArwe
16:35:00 <Zakim> -SteveBattle
16:35:02 <Arnaud> sandro, I'll clean up the minutes
16:35:07 <sandro> thanks.
16:35:18 <sandro> what do you mean?  aren't they already perfect?
16:35:20 <sandro> :-)
16:35:32 <Arnaud> :)
16:40:00 <Zakim> disconnecting the lone participant, pchampin, in SW_LDP()10:00AM
16:40:02 <Zakim> SW_LDP()10:00AM has ended
16:40:02 <Zakim> Attendees were pchampin, Arnaud, +1.845.454.aaaa, SteveS, JohnArwe, codyburleson, Sandro, Alexandre, SteveBattle, Ashok_Malhotra, bblfish, Roger, TallTed
16:50:41 <bblfish> bblfish has joined #ldp
18:09:27 <Arnaud> sandro, is commonscribe in limbo?
18:09:34 <Arnaud> 504 Gateway Time-out The server didn't respond in time.
18:09:38 <sandro> checking
18:10:01 <sandro> (it was fine during the meeting....)
18:11:31 <sandro> It was working for me, but a bit slow.   I restarted it, which may make it faster.
18:11:47 <Arnaud> :)
18:11:50 <Arnaud> now it works!
18:11:55 <Arnaud> thakns
18:11:57 <Arnaud> thanks
18:12:12 <sandro> Sure thing!
18:20:47 <bhyland> bhyland has joined #ldp
18:26:36 <bblfish> bblfish has joined #ldp
18:54:29 <SteveS> SteveS has joined #ldp
19:17:55 <Zakim> Zakim has left #ldp
20:25:20 <bblfish> bblfish has joined #ldp
23:34:38 <Arnaud> Arnaud has joined #ldp