Linked Data Platform (LDP) Working Group Teleconference -- 27 Aug 2012

<trackbot> Date: 27 August 2012

<oberger> don't get upset, no one's available ;)

<sandro> hmmm?

<Yves> no call this week, no?

<sandro> Yes call this week. There was an agenda, plus it was in the minutes of last week and it's in the channel title.....!

<Yves> ok, thanks (I don't see the channel title, and am lagging on email)

<Ruben> Hi all, I won't be able to participate on the phone (traveling), but I will be on IRC.

<bblfish> hi

<oberger> ouch

<oberger> hello Dr larsen

<oberger> SteveS: or from a plane

<oberger> ah

<oberger> Hi Arnaud

<bblfish> was there a conf call last week?

<oberger> summer time is hard time to stay focused

<sandro> Yes, bblfish

<bblfish> ah, I thought it was every two weeks in summer

<oberger> texting via REST APIs while driving ? booh

<sandro> scribe: SteveS

<scribe> scribenick: SteveS

Minutes from 22 August 2012

<oberger> nope

Arnaud asks if anyone has read minutes…no response

<bblfish> I still have to do my action item

<sandro> action-4?

<trackbot> ACTION-4 -- Steve Speicher to review SPARQL Graph Store Protocol and suggest how we should move forward with it -- due 2012-09-03 -- OPEN

<trackbot> http://www.w3.org/2012/ldp/track/actions/4

SteveS slacking on ACTION-4

SteveS to get to in next week or so

Action reviews

<oberger> reminder : no meeting next week (Labour day in the US)

bblfish will get to action-5

No update on ACTION-6, Michael had sent regrets for today

Upcoming meetings

Arnaud: Next telecon is in 2 weeks, September 10

sandro: looked into f2f room sizings for larger group, need WG members to actually register

…can get a bigger room if registration counts are high enough

sandro: recommends attending Wed sessions, many find it very valuable

<BartvanLeeuwen> :)

<ghard> Excuses for being late. Got network stability problems.

…costs about $50 USD / day for F2F

??? asked bblfish if room or something planned for f2f with WebID

<oberger> SteveS: oberger was me

<sandro> €45 until the close of registration (16 October 2012)

Use Cases and Requirements structure

<bblfish> So I am looking to organise a meeting for WebID and RWW Community Groups at TPAC Lyon, but we are still looking to see if we can get a room.

This is topic ACTION-9

<bblfish> But otherwise I will look to see if we can organise something anyway in Lyon.

Here's the proposal http://lists.w3.org/Archives/Public/public-ldp-wg/2012Aug/0137.html

Arnaud: positive comments from a number looking to hear if any other feedback

SteveBattle: has updated the wiki to take into account this change

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

namely changing Use Cases to User Stories, in prep for collecting use cases

<oberger> SteveS: /me added the link to the action in https://www.w3.org/2012/ldp/track/actions/9

…plan is to expand on the User Stories, then pull out the more detailed Use Cases and the more detailed Requirements

BartvanLeeuwen: Are the use cases/stories still open for contributions?

SteveBattle: Yes

<bblfish> good so I suppose my action item should be a user story

<bblfish> or integrated with one

SteveBattle: Assumes these will be very "CRUDy" based on the scope of this WG

<scribe> ACTION: SteveBattle to Create at least one Use Case by next meeting due 9/10/2012 [recorded in http://www.w3.org/2012/08/27-ldp-minutes.html#action01]

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

Arnaud: proposed to close action-9

close action-9

<trackbot> ACTION-9 Propose a new structure for the Use Cases and Requirements document closed

SteveBattle: SteveS and Michael (not on call) are interested in developing out the use cases more?

SteveS: Yes

Serialization format discussion and process

Arnaud: how does the process work for raising issues, like serialization format, and get reviewed and agreed on

sandro: someone raises (creates) an issue in the system, it goes into a pending state, then the WG decides if they accept the issue (something they will work on or valid) and it becomes open

…the WG then works on a resolution

Arnaud: believes that by definition of our charter and references to member submission, then we can start with that and open issues from there and get WG agreement

<RezaBFar> @Arnaud - sorry I'm late...

Arnaud: need to make sure the work we do with perhaps a more abstract model then we need to make sure we map it back to RDF model

<oberger> bblfish: ?

<bblfish> http://webid.info/spec/

bblfish: Expressed my thoughts on this and around the WebID spec. Did require a publish to N3 and clients understand both N3 and RDF/XML

<bblfish> GRDDL

<bblfish> XSPARQL

…would like to add that there is a requirement to support GRDDL for expressing what is supported

…it is not reasonable to convert some forms not into RDF

<bblfish> I think turtle is also a standard now http://www.w3.org/TR/turtle/

<RezaBFar> +1 to SteveBattle. That's exactly what I'm looking for.

SteveBattle: +1 for openness of serialization of RDF models, would like to support at least one. RDF/XML is reasonable as only W3C Rec spec, perhaps we have at least 1 minimum required

…server could expose what it supports

<AndyS> N3 goes beyond RDF - to be clear here, lets stick to the (soon to be) standard turtle.

<AndyS> N-triples will also be standardised by RDF-WG.

<Ruben> There's no advantage in this scenario to N3 instead of Turtle.

MacTed: says that Arnaud mentioned limit options, but believes we need to be more open to extensibility

<SteveBattle> That's interesting - I didn't know N3 went beyond RDF.

<Ruben> N3 adds quantification and variables.

<RezaBFar> So, I think there is 3 things: 1. Some standard query as part of the spec that provides format (so serialization format discovery) 2. Allowance for implementation of serialization of other formats. 3. Implementing at least 2 formats, 1 of which is RDF/XML, the other which we could agree on as a team.

MacTed: resource creation (POST) is that a server can chose to reject or handle that request as needed (transform or handle as is)

<oberger> MacTed: I didn't get your concern abouut RDF+XML...

<oberger> would like to see a transcript

<SteveBattle> What can't be encoded in RDF/XML?

<oberger> SteveBattle: that one yes

<AndyS> Properties with URIs that aren't representable as qnames e.g. http://example/1234

Arnaud: not against building an extensible spec but more on the point when there are too many options to achieve same/similar things

<AndyS> and http://example/property# and a few other cases. Tend to be unusual but can be tricky at scale if it gets into the data.

<oberger> RezaBFar: what problem ?

RezaBFar: Agrees with SteveBattle, doesn't see the need to limit to RDF/XML and have additional format

<oberger> verbosity ?

….XML processing has some overhead and should consider less-verbose options for efficiency

<oberger> JSON-LD would be very much appealing to all the Web2 devs

<oberger> YAML ? ;)

<sandro> oberger, more appealing than RDF/XML or than Turtle?

<oberger> sandro, people who do native JS ? probably

<MacTed> concern with RDF/XML: cannot serialize all RDF ... as AndyS says. see http://www.w3.org/DesignIssues/diagrams/n3/venn

<sandro> oberger, I was asking an (a) or (b) question, but I guess your answer is "both".

<oberger> sandro, sure, both : no extra parsing

<bblfish> So to summarise my point earlier:

<bblfish> 1. Turtle ( http://www.w3.org/TR/turtle/ ) or RDFxml perhaps a SHOULD ( on the http://webid.info/spec/ Turtle, RDFa, RDFxml were made MUST understand by client -- Turtle, RDFxml should be sent by server )

<bblfish> 2. Other formats should be supported but there should be a follow your nose to a GRDDL http://www.w3.org/TR/grddl/ so that one can automatically transforrm it to RDF graphs

<bblfish> 3. there is a complicating issue as this LDP requires PUT & POST which means that the server does need to be able to understand the formats sent.

<bblfish> 4. there will be some form of Graph query which will require that the format be queryable as a graph

ISSUE: Determine minimum serialization format for RDF data model

<trackbot> Created ISSUE-1 - Determine minimum serialization format for RDF data model ; please complete additional details at http://www.w3.org/2012/ldp/track/issues/1/edit .

<RezaBFar> Question: so now that it's an issue, what's next?

<RezaBFar> @Arnaud - how do we resolve the issue since it's a significant issue.

<AndyS> (caution - that diagram is confusion - Turtle can express any RDF - the reification point is not the point - it's old special syntax)

<oberger> we can move on when everybody's back from vacation more or less ?

<RezaBFar> @Arnaud - Makes sense. Thanks for the clarification.

Arnaud: answering RezaBFar, need to have a proposal and reach consensus on the issue and record it

sandro: can be good collect more facts on the matter and may need to get to point where just need to pick one

<RezaBFar> @sandro - thanks.

<RezaBFar> Sandro's opinion?

<oberger> next in 2 weeks ?

<RezaBFar> on RDF/XML, etc.? Sandro was saying something?

Arnaud: any other agenda items?

<oberger> sandro, my suggestion on JSON-LD is purely speculative

sandro: is a big win in mental clarity on a simple format but think still open in my mind if JSON-LD meets that
... if I had to pick something right now, it would be Turtle but may be things I'm forgetting about RDF/XML

<BartvanLeeuwen> +1 to bblfish remarks to JSON-LD remark

bblfish: Wonder if there is anything around efficiency with XML binary formats that apply with RDF/XML

<RezaBFar> FWIW, I agree with Sandro on Turtle.

<MacTed> ... s/JSON-GRDDL/JSON-LD/

bblfish: question to JSON people, when you publish in json-grddl can be represented as any tree as root of graph and could it confuse them?

MacTed: was in JSON-LD group, started as a way to get JSON people who are using structured data in a more interoperable ways based on RDF concepts
... for some similar implementations Turtle makes it very easy but lacks some of the RDF concepts, as linked to before

<oberger> Turtle is great to read by humans, discovering specs, but in all languages ?

<bblfish> +1 for Turtle as base, easier to understand

<sandro> STRAWPOLL: Which one RDF serialization should we pick as our 1-required-serialization, if we had to pick one today? (eg rdf/xml, turtle, json-ld, RDFa, or NONE)

<AndyS> Turtle can express any RDF.

<sandro> turtle

<RezaBFar> Turtle

<bblfish> +1 Turtle

<BartvanLeeuwen> Turtle

<oberger> RDF+XML

<MacTed> Turtle

<AndyS> Turtle

<ArthurK> turtle

<oberger> for implementation

sandro: straw man poll of preferred RDF serialization

<SteveBattle> (turtle and rdf/XML)

Turtle

<Kalpa> rdf/XML

<RezaBFar> So, I meant (Turtle and RDF/XML)

<MacTed> AndyS - any chance you can rework that Venn with current status? possibly including JSON-LD?

<MacTed> there may be other serializations that didn't exist when it was drawn that should also be added in...

<bblfish> well rdf/xml would be a SHOULD

<oberger> which libs are turtle compatible ?

<Zakim> sandro, you wanted to propose strawpoll

<Arnaud> andy?

AndyS: found the diagram was someones opinion for a long time ago and not right

sando: agrees

MacTed: would like to see it updated

<oberger> bye

<SteveBattle> bye

<BartvanLeeuwen> thanks and bye

<bblfish> bye

<ghard> bye

<Kalpa> bye

<sandro> MacTed, note that http://www.w3.org/DesignIssues/diagrams/n3/venn is about expressibility, not RDF. Full N3 and SPARQL WHERE include lots of things that are not, and never will be, RDF.

rssagent, generate minutes

- DRAFT -

Linked Data Platform (LDP) Working Group Teleconference

27 Aug 2012

Attendees

Contents