Linked Data Platform (LDP) Working Group Teleconference

17 Sep 2012

See also: IRC log




<trackbot> Date: 17 September 2012

<bblfish> hi

<svillata> Zackin, ??P27 is me

<svillata> Zakin, ??P27 is me

<scribe> scribe: johnarwe

arnaud: approval of minutes from last week

<SteveS> +1 on minutes from Sept 10

arnaud: objections to approving as stated?

<Yves> +1

approved w/o objection

arnaud: update wiki page with F2F attendance/not

<bblfish> I added myself and paid.

arnaud: room size based on registered attendees, so in our interest to pre-reg

action items

arnaud: last week closed issue on serialization, created action on editors to draft as decided
... no open issues.

sparql graph store protocol, overview by steve speicher

steves: expose graph store via http protocol
... stated goals different
... difficult to align if one starts from stated goals
... OTOH, based on submission, possible that LDP impls could use graph stores as back end(s)
... GSP requires 1 of 3 formats if request omits Accept header. LDP requires Turtle only, so gap.

<mhausenblas> Michael: Sorry for being late, prev meeting took longer than expected

<bblfish> this I suppose http://www.w3.org/TR/sparql11-http-rdf-update/

steves: GSP wording of POST says it's a merge, LDP says it depends on the resource - BP resource = undefined, BP container = creating new resource and adding new resource's URI to the container
... POSTing to graph store URI = add new graph to graph store, at uri assigned by GS. No definition of semantics of GET on GS URI.
... "squinting" says some similarities at least.

ericp: sparql WG issue - does saying that POST means append "force the hand" of implementations not claiming to be a GS.
... e.g. append could be notional, whole bunch of triples show up (new membership in container, new resource) others change (last page != nil). *Could* say that GSP applies to resources created (e.g. asset in submission ex), others LDBP containers, and GSP does not control what happens in LDBP containers. Only way I can see claiming conformance to both in single impl.

<Zakim> ericP, you wanted to ask if squinting would get us test-case compatibility

who is speaking about sparql impl?

ashok: should we be thinking about how to align them?

ericp: could parameterize GSP, so there's an append type of resource

arnaud: GSP is in Last Call, what is sense from WG on whether it is likely to proceed vs WG being eager to make changes, [implication: as alignment might require]?

sandro: no time to waste, expect Rec by end of year. WG added language to narrow scope down to SPARQL only, not general HTTP access to RDF. If LDP WG thinks GSP is harmful to the Web, LDP would have to object. Spec might well die if that happened.

dret: any problem separating them cleanly from a rest perspective, so an implementation could offer one link for each protocol? that the way to do it from REST perspective. "Home" resource, "dispatch" with 1 link for GSP, 1 for LDP.

<Zakim> ericP, you wanted to talk about sparql service description

sando: could work

ericp: service desc could differentiate the resources it owns/manages. Not sure if I can say I have a GS if only some of its resources adhere to GSP.

<Zakim> bblfish, you wanted to comment on home page

<dret> general home document work: https://datatracker.ietf.org/doc/draft-nottingham-json-home/ (probably will be extended to cover XML as well)

bblfish: (henry story) home page work reminds one of atom's service page discussion, resource should describe itself.

sandro: that is how sparql serv desc works, but that only applies to sparql endpt not to member resources

henry: link header for metadata?

<dret> Atom(Pub) can use 'service' links so that any interaction with a Atom feed or entry can link to the AtomPub service document.

sandro: would work, no one mandating it tho. sparql WG chose not to define that, they wanted something simple and this aspect was an add-on they had little experience with, not core to their work.

<dret> resources could link to both service documents, if they are served by a container implementing two access models. clients then could choose which one to follow.

<sandro> technical solution could just be providing RDF type information for the resource, using an HTTP Link header with an rdf:type property.

steves: LDP container proposal was one of several. did discuss link header(s) too, open if we wanted to change it. not sure what it would look like if mixed with GSP. Could highlight as indep way to access same resources, could define a mapping from LDP to GSP, could make some changes in LDP spec to improve alignment, perhaps recommend how POST works if an impl wants to offer both protocols.

sandro: this WG could suggest they add headers/whatever, w/in LDP's scope

arnaud: any opinions on whether or not LDP should make any comments on GSP?

<sandro> AndyS, you here?

<bblfish> perhaps we should leave this issue open. I need to look at it more carefully myself to understand the issue

sandro: no strong need to hold GSP up
... others might see such a need, if so should speak up

steves: no reason to hold them up

<bblfish> perhaps just put a summary on the wiki

<bblfish> so we can refer to it

arnaud: thanks steve for investigating. fulfilled your part.

<scribe> ACTION: arnaud to add topic on wiki on LDP vs GSP on home page [recorded in http://www.w3.org/2012/09/17-ldp-minutes.html#action01]

<trackbot> Created ACTION-13 - Add topic on wiki on LDP vs GSP on home page [on Arnaud Le Hors - due 2012-09-24].

<bblfish> wow ACTION 13 :-)

* touch typists are dangerous ;-)

trackbot, close action-4

<trackbot> ACTION-4 Review SPARQL Graph Store Protocol and suggest how we should move forward with it closed

bblfish: consent to close action, can always re-open later if new material comes up

use case document

mike and ram working on document

spec draft

steves: working on tool setup, pretty much done. will not make any changes aside from things required to fit it into the template.

arnaud: you reviewed timeline, ok with it?

steves: y

WG time line

<mhausenblas> +1 to the timeline

<dret> i will add it to the wiki

arnaud: chairs took action last time to create time line, went through the right to left schedule exercise, posted.


<dret> give me a minute, i'll update it now that we have approved it

need to update that link

steves: tracker is not linking in email activity, e.g. on 11 I do not see discussion threads. is that a setup problem?

trackbot, close action-11

<trackbot> ACTION-11 Define a timeline for working on the editor's draft and the FPWD (chairs action, i.e. Arnaud and dret) closed

action on yves to Look at tracker config to ensure when an issue is mentioned in an email that the action is updated with the email thread.

<trackbot> Sorry, couldn't find user - on

action yves to Look at tracker config to ensure when an issue is mentioned in an email that the action is updated with the email thread.

<trackbot> Created ACTION-14 - Look at tracker config to ensure when an issue is mentioned in an email that the action is updated with the email thread. [on Yves Lafon - due 2012-09-24].

back to use cases and requirements document

arnaud: steve battle on today; WG looked at content as of last week, liked direction

<bblfish> do we have a link?

battle: got some feedback from steves, working on terminology consistency, took first user story (social contact infor), extracted use case (GET), raised some issues

<Arnaud> http://www.w3.org/2012/ldp/wiki/Use_Cases_And_Requirements

<bblfish> perhaps this in particular: http://www.w3.org/2012/ldp/wiki/Use_Cases_And_Requirements#Use_Cases ?

<dret> http://www.w3.org/2012/ldp/wiki/Main_Page#Timeline is updated now (ACTION-11)

battle: need to clarify if resources are things or representations of things, i.e. do we need 303 redirect or not

arnaud: as you work on doc, as you see issues, please raise them into tracker system (best) or email list.

<dret> if there's no discussion on the email list, then we see them during the call and will address open issues and actions in the meeting.

arnaud: unrealistic perhaps to have public working draft by end of Oct for f2f, but would like at least an editor's draft.

battle: imp thing is to label/number use cases

<Zakim> bblfish, you wanted to comment on TPAC question

<bblfish> bblfish: yes, call for sponsorship for WebID and RWW Community group for 1 day $1000

<SteveS> When will our TPAC agenda be drafted/set? Be good to reserve time for UCR finalizing and spec issue resolution

<ericP> €1000 instead of $1000, iirc

<ericP> (from what Henry said)

<bblfish> ah yes. thanks :-)

<bblfish> 1352.74 euros exactly

<bblfish> todo: Will send mail to mailing list on the topic of webid and ldp

Summary of Action Items

[NEW] ACTION: arnaud to add topic on wiki on LDP vs GSP on home page [recorded in http://www.w3.org/2012/09/17-ldp-minutes.html#action01]
[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.136 (CVS log)
$Date: 2012/09/17 15:02:07 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.136  of Date: 2011/05/12 12:01:43  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

Succeeded: s/speicer/speicher/
Succeeded: s/ericp:/sandro:/
Succeeded: s/ ot / to /
Succeeded: s/them/open issues and actions/
Found Scribe: johnarwe
Inferring ScribeNick: JohnArwe

WARNING: No "Present: ... " found!
Possibly Present: AndyS Arnaud Ashok_Malhotra BartvanLeeuwen IBM IPcaller JohnArwe MacTed Michael OpenLink_Software P12 P13 P27 P6 Ruben Sandro Yves aabb aacc ahaller2 antonis ashok battle bblfish betehess cygri dret ericp henry joined kalpa krp ldp mhausenblas nmihindu oberger rgarcia sando stevebattle steves svillata todo trackbot
You can indicate people for the Present list like this:
        <dbooth> Present: dbooth jonathan mary
        <dbooth> Present+ amy

WARNING: No meeting chair found!
You should specify the meeting chair like this:
<dbooth> Chair: dbooth

Found Date: 17 Sep 2012
Guessing minutes URL: http://www.w3.org/2012/09/17-ldp-minutes.html
People with action items: arnaud

[End of scribe.perl diagnostic output]