13:56:29 <RRSAgent> RRSAgent has joined #ldp
13:56:29 <RRSAgent> logging to http://www.w3.org/2013/04/15-ldp-irc
13:56:31 <trackbot> RRSAgent, make logs public
13:56:31 <Zakim> Zakim has joined #ldp
13:56:33 <trackbot> Zakim, this will be LDP
13:56:33 <Zakim> ok, trackbot; I see SW_LDP()10:00AM scheduled to start in 4 minutes
13:56:34 <trackbot> Meeting: Linked Data Platform (LDP) Working Group Teleconference
13:56:34 <trackbot> Date: 15 April 2013
13:56:39 <pchampin> pchampin has joined #ldp
13:57:11 <Zakim> SW_LDP()10:00AM has now started
13:57:32 <Zakim> +??P2
13:57:48 <pchampin> zakim, ??P2 is me
13:57:48 <Zakim> +pchampin; got it
13:58:41 <Zakim> +Sandro
14:00:13 <Zakim> + +1.661.748.aaaa
14:00:15 <cody> cody has joined #ldp
14:00:16 <mielvds> mielvds has joined #ldp
14:00:26 <Zakim> - +1.661.748.aaaa
14:00:34 <Zakim> +Arnaud
14:00:35 <Zakim> +JohnArwe
14:00:37 <JohnArwe> regrets: Bart, Andy (likely)
14:00:41 <Zakim> +[IBM]
14:00:55 <SteveS> zakim, [IBM] is me
14:00:55 <Zakim> +SteveS; got it
14:01:41 <Zakim> +Ashok_Malhotra
14:02:34 <Zakim> +[IPcaller]
14:02:37 <Zakim> + +329331aabb
14:02:44 <rgarcia> rgarcia has joined #ldp
14:02:51 <cody> Zakim, IPcaller is me
14:02:51 <Zakim> +cody; got it
14:02:51 <mielvds> zakim, +329331aabb is me
14:02:52 <Zakim> +mielvds; got it
14:02:55 <Arnaud> zakim, who is on the phone?
14:02:55 <Zakim> On the phone I see pchampin, Sandro, Arnaud, JohnArwe, SteveS, Ashok_Malhotra, cody, mielvds
14:03:07 <SteveS> sandro, you had mentioned hacking together some text for a patch document proposal…did you ever get to that?
14:03:36 <Zakim> +??P7
14:03:52 <rgarcia> zakim, ??P7 is me
14:03:52 <Zakim> +rgarcia; got it
14:04:05 <Zakim> +OpenLink_Software
14:04:17 <TallTed> Zakim, OpenLink_Software is temporarily me
14:04:17 <Zakim> +TallTed; got it
14:04:20 <TallTed> Zakim, mute me
14:04:20 <Zakim> TallTed should now be muted
14:04:41 <rgarcia> I'm having problems listening today
14:04:44 <rgarcia> I'm sory
14:04:51 <rgarcia> s/sory/sorry/
14:05:12 <TallTed> Zakim, unmute me
14:05:12 <Zakim> TallTed should no longer be muted
14:05:14 <Zakim> +SteveBattle
14:05:14 <sandro> mielvds, who are you....?
14:05:16 <cody> I can give it a go. Of course, I'm not the most practiced Scribe.
14:05:22 <mielvds> maybe I should introduce myself
14:05:27 <TallTed> Zakim, mute me
14:05:27 <Zakim> TallTed should now be muted
14:05:41 <cody> scribe: cody
14:05:41 <Arnaud> scribe: cody
14:05:45 <mielvds> mielvds has joined #ldp
14:06:13 <cody> Arnaud: Announcement on "CommonScribe"
14:06:32 <sandro> https://www.w3.org/2013/meeting/
14:06:43 <sandro> https://www.w3.org/2013/meeting/ldp/
14:06:53 <sandro> https://www.w3.org/2013/meeting/ldp/2013-04-15
14:07:10 <cody> ??: At that URL are all the IRC channels. Go directly to LDP by clicking on that. You will see all the days. At the top you will see the day. And you will see the log is only about a second behind IRC.
14:07:29 <JohnArwe> s/??/Sandro/
14:07:31 <cody> s/??/snadro
14:07:44 <cody> ?/snadro/Sandro
14:08:07 <cody> Sandro: There may be a lot of new bugs; it's new code.
14:08:07 <mielvds> mielvds has joined #ldp
14:08:20 <cody> … just let me know if you find anything buggy.
14:08:43 <sandro> mielvds, can you type those names on IRC?
14:08:59 <sandro> Ah, got it Vander Sande, Miel
14:09:15 <cody> ??: My name is ??, and I am joint this group to divide some of the work Reuben has. I do not have much exp with working groups, but hope to get up to speed fast. I do have Semantic Web systems experience.
14:09:36 <JohnArwe> s/??:/Miel:/
14:09:53 <roger> roger has joined #ldp
14:10:43 <cody> Arnaud, hearing no objections, I hereby approve meeting minutes of April 8.
14:10:53 <Zakim> + +44.208.573.aacc
14:11:02 <JohnArwe> RESOLUTION: meeting minutes of April 8 approved
14:11:10 <cody> … Raul, could you provide some information on the venue, hotels, etc.?
14:11:11 <sandro> topic: Next F2F
14:11:37 <cody> Raul: OK, we'll try to provide something in the wiki by next week.
14:11:42 <sandro> Arnaud: please update the wiki page
14:11:45 <JohnArwe> http://www.w3.org/2012/ldp/wiki/F2F3
14:11:50 <cody> topic: Actions and Issues
14:11:56 <Zakim> +[GVoice]
14:12:01 <ericP> Zakim, [GVoice] is me
14:12:01 <Zakim> +ericP; got it
14:12:24 <Zakim> +bblfish
14:12:54 <cody> Arnaud: Suggest we close Action 44. Objections?
14:13:03 <bblfish> bblfish has joined #ldp
14:13:11 <bblfish> hi
14:13:15 <cody> … if you don't agree with the change that was made, please bring it up.
14:14:26 <cody> … We have a lot of open actions and I am concerned that there was not much open discussion this week. I really need everybody to stay engaged. We have to tackle this open actions and issues to stay on track. I would appreciate if we upped our efforts.
14:14:51 <cody> ??: Minor update: I have action 45 in progress.
14:16:30 <JohnArwe> s/??:/SteveSpeicher:/
14:17:21 <cody> Arnaud: We are supposed to publish a note indicating that we are not defining requirements for Access Control within scope of the spec. Some of us felt that including it would be a rabbit (rat?) hole that would hold us down for a long time. So, instead, we decided to publish a note discussing some use cases and requirements. Ashok has taken the lead in drafting the first draft of this note. Ashok, you asked for some time on the agenda?
14:17:56 <cody> Ashok: I sent a draft in email; something I thought we could start speaking about.
14:18:46 <bblfish> q+
14:18:49 <cody> … I got a couple of replies. One from Andy. One from Pierre. Their comment was basically that the Access Control was going to be provided, not by the database, but by a layer on top of it. I didn't really understand that. I'd like clarification.
14:20:01 <cody> Pierre: Yeah, um. My experience with web applications is that they usually do not rely on access control from the database, but it's at the application layer itself.
14:20:29 <cody> … In my experience, it's not the case with current web applications. I would expect that LDP apps follow the same path.
14:20:56 <cody> … I think it is just misleading to say that it would be done in the storage level.
14:21:08 <ericP> i think our mission is to explore some likely ways that the access control *might* be implemented
14:21:31 <TallTed> Zakim, who's noisy?
14:21:37 <cody> Ashok: asking a question (scribe missed)
14:21:43 <Zakim> TallTed, listening for 10 seconds I heard sound from the following: pchampin (94%)
14:21:48 <ericP> and to make sure that we don't explicitly preclude those approaches
14:22:20 <cody> Pierre: I'm just basing this on what I see on current applications.
14:22:50 <bblfish> q?
14:22:51 <cody> … All the tables in the app are usually accessed by a single user (db user)
14:23:06 <Arnaud> ack bblfish
14:23:06 <cody> … and the app controls the security for each authenticated user.
14:23:40 <ericP> q+ to say that there probably enough apps which use the underlying database both to serve content and support ACLs that it's worth exploring those use cases
14:23:52 <rgarcia> rgarcia has joined #ldp
14:25:01 <Zakim> -ericP
14:25:06 <cody> Henry: There's non need to be talking about databases and file systems and such. What's important is what's communicated. Web Access control is a very good example of the kind of architecture I was sort of thinking about.
14:25:15 <Arnaud> ack eric
14:25:15 <Zakim> ericP, you wanted to say that there probably enough apps which use the underlying database both to serve content and support ACLs that it's worth exploring those use cases
14:25:18 <Zakim> +[GVoice]
14:25:27 <ericP> Zakim, [GVoice] is me
14:25:27 <Zakim> +ericP; got it
14:25:29 <cody> Ashok: If people have various examples like that, which I could go look at - that would be a big help to me.
14:27:51 <cody> ericP: Our mission is to note possible mechanisms for access control, so that we don't fail to provide the appropriate hook(s). In the places where I have done this, the same db that holds the resources also holds the users and acls. If we want to address these use cases, out's worth writing out the various approaches; there could be several.
14:28:20 <cody> s/out's/it's/
14:29:35 <cody> Arnaud: Ashok made the assertion that Access Control is provided by the underlying storage. That's an implementation detail.
14:31:19 <cody> Arnaud: I would expect us to specify the kind of access controls we have, but not how to do it. We have to explain the useful operation of the LDP spec and the kind of access control expected. E.g. identification of the user, read and write at the resource level...
14:31:57 <TallTed> Zakim, unmute me
14:31:57 <Zakim> TallTed should no longer be muted
14:32:52 <cody> TallTed: I would suggest that we describe this stuff with the decoupling clear. The description of any use case should be based on the granular functionality.
14:33:17 <cody> ericP: The purpose of writing use cases is to determine if that changes anything (in the spec).
14:34:15 <bblfish> q+
14:34:26 <Arnaud> ack bblfish
14:34:48 <cody> scribe asks ericP so note that last salient point himself.
14:36:01 <JohnArwe> s/so note/to note/
14:36:21 <cody> Ashok: Henry, if we end up writing an RDF specification of access control, then we could start to think about how that could or couldn't be tacked on to various existing AC mechanisms?
14:37:02 <ericP> ericP: if we explore use cases where a database provides both the content and the ACLs, as well as the decoupled cases, we can see they produce any different requirements
14:37:25 <ericP> ... ideally, they won't, in which case we can factor out the coupled case.
14:37:30 <ericP> .
14:38:03 <cody> Henry: GET, PUT, POST, DELETE will either be allowed or denied. The important thing is the structure of the access control on the action - not the database.
14:38:33 <cody> Henry: Who gets access to the resource could be specified in RDF.  Authentication could be something else (OpenID, whatever)
14:39:06 <cody> Arnaud: Let's move on. Advise everyone to continue with this thread that Ashok has started. And there is quite a biot of material on the wiki page.
14:39:14 <cody> topic: ISSUES
14:40:03 <cody> Arnaud: We did not close any of the issues we discussed last week. Seems like some people were not quite up to date. Hoipefully you used the week to catch up and leo, lets see if we can close one or two.
14:40:08 <cody> topic ISSUE 58
14:40:34 <cody> TOPIC: ISSUE 58
14:41:41 <cody> TallTed explained his last email to ldp-wg
14:41:45 <JohnArwe> http://lists.w3.org/Archives/Public/public-ldp-wg/2013Apr/0048.html is the email
14:41:59 <cody> Arnaud:
14:42:05 <bhyland> bhyland has joined #ldp
14:42:10 <pchampin> q+
14:42:13 <bblfish> q+
14:42:14 <TallTed> you can see here -- http://www.w3.org/2012/ldp/track/issues/58
14:42:35 <Arnaud> ack pchampin
14:43:15 <cody> ??: I think another option was raised: to put information in the header.
14:43:40 <JohnArwe> s/??:/Pierre-Antoine:/
14:43:44 <SteveS> q+
14:43:57 <cody> Arnaud: What do we put in the content in the RDF versus what we put in the HTTP header. It doesn't seem to be clear cut to put it where. This is the question.
14:44:04 <Arnaud> ack bblfish
14:45:39 <cody> Henry explains a concern, which he also emailed this morning.
14:46:01 <cody> TallTed: I, as a server, have given you every attribute which I know.
14:46:15 <pchampin> q+
14:47:05 <bblfish> <> rdf:member <mbr> .
14:47:19 <Arnaud> ack steves
14:47:27 <bblfish> <> log:contains <mbr> .
14:47:59 <cody> SteveS: I can think of pros and cons to a large number of different ways of doing it.
14:48:20 <Arnaud> ack pchampin
14:48:28 <stevebattle> q+
14:48:29 <cody> Arnaud: We went to having nothing, no flag, to having one on every resource. Maybe we shouldn't overshoot it.
14:49:23 <cody> ??: Is it container wide or resource wide? If the probe;em is taken that way - it sounds that it is related to the protocol and for me pushes the balance towards the information being in the header.
14:49:44 <cody> s/probe;em/problem/
14:50:02 <JohnArwe> s/??:/pchampin/
14:50:11 <roger> +q
14:51:16 <Arnaud> ack steveb
14:51:50 <roger> q-
14:51:54 <cody> Scribe missed pchampin's last point; type it in please if you could.
14:52:03 <roger> -q
14:52:48 <pchampin> pchampin: is there a way in HTTP to tell "this description I have you for URI1, you can cache it under URI2"
14:53:04 <pchampin> ... this would give us resource-level control for free
14:53:19 <cody> TallTed: That's the point of this extra attribute. The server can tell you - 'I've done this thing.' I'm getting on a container. Does the server tell you I have fully described every resource on this container? That I have described these, but not these others? Or does it tell you nothing?
14:54:17 <cody> … In the middle option, there does not need to be a GET for description on every resource; only those for which the server has already said is not complete
14:55:00 <cody> Arnaud: I think Richard's proposal to add one level of control. Not as fine grained as what we could do. If someone wants to make the case for finer level of control, they should do so.
14:55:06 <roger> +q
14:55:13 <cody> TallTed: That's what I am doing - making that case.
14:55:49 <Arnaud> ack roger
14:55:51 <stevebattle> I think if you're going to do it at all, you may as well go with Ted's fine-grained approach
14:55:56 <cody> Arnaud: OK - does anybody disagree with Ted's point? Because Ted is making a strong case for more detail level that Richard's proposal.
14:56:02 <SteveS> I see the possible need for per member resource but not urgent need for me in 1.0
14:56:03 <cody> Roger: I agree with Ted.
14:56:44 <Arnaud> Proposed: accept as a requirement the need to indicate which member resources are fully inlined
14:57:12 <cody> Arnaud: I propose that we need to indicate which resources are fully inlined.
14:57:15 <TallTed> +1
14:57:15 <roger> +1
14:57:19 <stevebattle> 0
14:57:25 <SteveS> 0
14:57:28 <rgarcia> 0
14:57:34 <pchampin> +.5
14:57:40 <JohnArwe> +0 (see the general usefulness, not convinced it's meets-min for LDP)
14:57:42 <cody> 0
14:58:27 <mielvds> mielvds has joined #ldp
14:58:30 <Arnaud> Resolved: accept as a requirement the need to indicate which member resources are fully inlined
14:59:12 <cody> Arnaud: For next week, it would be good if one of you could update the proposal on how we do this. Given that requirement, we cannot take Richard's proposal.
14:59:22 <SteveS> I think it would be interesting if a solution was in place that indicated the ETag, such as <member-resource> ldp:resourceInlined "etag value".  This would be useful information, obviously for caching cases
14:59:44 <cody> … Ted, since you feel strongly, maybe you can post proposal to the mailing list?
14:59:51 <cody> TallTed: OK
15:00:03 <bblfish> Issue-14?
15:00:03 <trackbot> ISSUE-14 -- Include clarifications about ordering in BPC representations -- open
15:00:03 <trackbot> http://www.w3.org/2012/ldp/track/issues/14
15:00:04 <cody> TOPIC: ISSUE 14
15:00:29 <cody> Arnaud: We are out of time, but let's just scrub over this issue quickly.
15:01:03 <cody> TallTed: I apologize, but I have not been able to make progress on that last week.
15:02:39 <cody> Arnaud: My role, as chair, is to push us forward. I cannot afford to have all these issues open without action. Next week, I expect us to make a decision on this.
15:02:47 <Zakim> -Ashok_Malhotra
15:02:55 <cody> Arnaud: Time's up. Thank you all!
15:02:56 <Zakim> -bblfish
15:02:57 <Zakim> -SteveS
15:02:59 <Zakim> -TallTed
15:03:00 <Zakim> - +44.208.573.aacc
15:03:01 <Zakim> -mielvds
15:03:05 <Zakim> -rgarcia
15:03:07 <Zakim> -Arnaud
15:03:08 <Zakim> -SteveBattle
15:03:24 <cody> cody has left #ldp
15:03:28 <Zakim> -cody
15:03:32 <Zakim> -pchampin
15:04:25 <Zakim> -JohnArwe
15:05:34 <Zakim> -Sandro
15:05:39 <Zakim> -ericP
15:05:41 <Zakim> SW_LDP()10:00AM has ended
15:05:41 <Zakim> Attendees were pchampin, Sandro, +1.661.748.aaaa, Arnaud, JohnArwe, SteveS, Ashok_Malhotra, cody, mielvds, rgarcia, TallTed, SteveBattle, +44.208.573.aacc, ericP, bblfish
15:12:39 <davidwood1> davidwood1 has joined #ldp
15:43:42 <davidwood> davidwood has joined #ldp
17:06:40 <bhyland> bhyland has joined #ldp
17:19:26 <Zakim> Zakim has left #ldp