Linked Data Platform (LDP) Working Group Teleconference -- 14 Oct 2013

<trackbot> Date: 14 October 2013

<scribe> scribe: Alexandre

<scribe> scribenick: betehess

you guys need to ask the team contact to put the publication on /TR

Arnaud: let's start with the minutes of last meeting
... can be approved?
... there were some resolutions
... hope we captured them well

<stevebattle6> no objections from me.

APPROVED: minutes from last week

Arnaud: is that a holyday?
... hope it won't be a holyday next week, we'll have a call
... are there any actions we can dispose of somehow?
... we don't have too many
... getting better at it
... let's go with agenda
... last time, JohnArwe pointed it us at a diff in Mercurial

<bblfish> hi

Arnaud: would like to get this officially this in

<Arnaud> Proposal: Change following to informative (re: redefining HTTP), diff: http://dvcs.w3.org/hg/ldpwg/rev/eeff2a51723d

<stevebattle6> +1

Arnaud: people should have had time to review it

<JohnArwe> +1

<BartvanLeeuwen> +0

<JohnArwe> (Steve Speicher having IRC trouble ... voice +1)

betehess: what about hash-uris

JohnArwe: just copied from before. anybody who knows HTTP knows not send the hash part

<JohnArwe> betehess question was on section 5.2.6

Arnaud: we're not redefining HTTP

<bblfish> that remark seems to stem from the problem arising out of confusing LDPC members that are LDPRs ( resources that are docs without a hash ) and the fact that some people want members with hash uris. That is still why I think one should put it in terms of ldp:created

<Arnaud> Resolved: Change following to informative (re: redefining HTTP), diff: http://dvcs.w3.org/hg/ldpwg/rev/eeff2a51723d

<ericP> +1 to the spirit, though i've not done my homework

Arnaud: next topic is the straw poll
... about to use those two kind of servers
... chocolate vs vanilla?
... it's just terminology

<SteveS> +1

Arnaud: how to define the two modes in the spec
... do people have an opinion to help the editors?
... must decide between SHOULD and MUST
... some forms are more compact than others

<BartvanLeeuwen> +q

<JohnArwe> http://lists.w3.org/Archives/Public/public-ldp-wg/2013Oct/0040.html

BartvanLeeuwen: what seems unclear to me is where you're allowed to @@@
... with the spec right now, was confused

JohnArwe: if you look at the draft, I've given more information
... making difference between the server and the client
... the client only sees the interface
... you don't have to obey all rules when you serve an image for example (it's not an LDPR)
... if the client doesn't know, it's just plain HTTP

<ericP> JohnArwe: an LDP server MUST adhere to the defined LDPR interface when serving an LDPR. (not when serving other docs)

BartvanLeeuwen: not exactly what I meant
... had the impression that I could mix vaniall and chocolate

Arnaud: I think there is no mixing, it's either chocolate or vanilla

BartvanLeeuwen: had the impression it wasn't clear

Arnaud: if there is anywhere a MUST that you don't make, then you end up being chocolate

<ericP> +1 to 1

Arnaud: do people have preference for editorial style?

<ericP> i find 1 the easiest to scan

Arnaud: I prefer 1

<JohnArwe> sandro is not here, but offline he said +0/+1/+1

+1 to 1

scribe: minimizes the duplication, makes it easier to read

bblfish: you need to make the different between vanilla and chocolate
... eg. POSTing anythign to LDPC unless there is a restriction
... not sure there is a difference there, it's a rule that applies to both, right?

Arnaud: people are still trying to understand the difference between the two
... as of now, we have a proposal to make changes, introducing those two modes
... emphasizing the MUST/SHOULD
... it's editorial style for when we need to deal with MUST and SHOULD

<SteveS> +1 to 1

<bblfish> my suggestion: LDP servers MUST xxx unless a restriction on the LDPC of R is stated in which case ...

Arnaud: the editors could have done than on their own, but they ask for input

<krp> +1 to 2 (weak preference)

ericP: looks like people are going for proposal 1

<krp> trying to unmute :)

<bblfish> I think these ways of putting things do not make clear how a client would know he is dealing with one server or the other

<bblfish> -1

<stevebattle6> +1 to option 1

<krp> but just due to preference of language

<BartvanLeeuwen> +1 option 1

<bblfish> that is in my view the vanilla/chocolate is an a posteriori distinction.

Arnaud: it's not like we need to make a resolution
... let's move with the agenda
... would like to speak about LC comments
... not sure we're handling them well
... would like the team to discuss with timbl

<JohnArwe> The tracker's output does not seem so useful today.

Arnaud: see if the changes are ok with him
... would like to save us another LC
... for the other comments, we have a tracker

<JohnArwe> Most tracker output lines start with (for me at least): Warning: mb_detect_encoding() expects parameter 1 to be string,

Arnaud: to keep track of them and respond to them
... commenters have a period to say if they agree with our responses
... not sure we're doing a good job so far
... looks like the editors are in sync there
... issue 81

http://www.w3.org/2012/ldp/track/issues/81

scribe: about membership predicates names
... but people have been ignoring my poll

<bblfish> we are hearing some typing I think

scribe: so I put it on this agenda
... it's different from the description of the issue

steveS: must include pierre-antoine's comment

Arnaud: naming is a challenge

bblfish: there may be another possibility
... don't think it was added as an option
... maybe was too radical?

<JohnArwe> something funky with the doodle survey too. I definitely voted (even mentioned +0s in comments), but it's not showing up here.

steves: I've put forward what dwoods put together

Arnaud: don't think anything was intentional
... bblfish, you can add that to the wiki

bblfish: ok

<bblfish> where is the wiki again?

<SteveS> http://www.w3.org/2012/ldp/wiki/ISSUE-81

Arnaud: can you do that now?
... bblfish, please add that to wiki and we can postpone to next week
... is there an issue with timezones and not having time to answer?

ericP: had questions as well

[people discussing meaning of proposals]

ericP: aaaaaaouwww

BartvanLeeuwen: was everything coming from last f2f?
... was missing references to the minutes

Arnaud: that's the case
... look at issue 81
... it's written there
... what we call memberSubject is the object for the membership inverse

<JohnArwe> last F2f minutes (day 1) http://www.w3.org/2013/meeting/ldp/2013-09-12 (day 2) http://www.w3.org/2013/meeting/ldp/2013-09-13

Arnaud: led to people saying that the spec is confusing
... added membershipInverse, and @@, and those additions made everything bogus

<JohnArwe> w3c server is serving them "slowly" now for some reason.

BartvanLeeuwen: will take time to review before next week

bblfish: my proposal is a structural one
... was interested in quick straw poll to understand what people want

steves: the proposal is about the words
... if structural proposal just add it there

bblfish: ok, I'll take that into account

Arnaud: people, please respond to the polls
... let's not talk about the PATCH
... even though we want a solution, I don't see it coming

ericP: I remember sandro sending mail
... suggesting an approach
... as we had conflicting proposals
... implementation feedback would help move forward

Arnaud: I think there are other proposals being sent
... my fear there will be more than 2, more like 3 or 4

betehess: what would it take to work on patch

Arnaud: I've seen too many proposals
... the problem is time
... the feature is at risk
... a ML was set up

<bblfish> what was the mailing list?

Arnaud: people should work there to sort it out

bblfish, public-ldp-patch@w3.org

<bblfish> thanks

Arnaud: back to chocolate vs vanilla
... answered by pierre-antoine
... wish there would be more discussions

JohnArwe: you're confusing 2 threads
... only roger answered to that thread
... didn't appear to make concrete proposals

Arnaud: what do people think?
... is there a real need for this vanilla vs chocolate thing?
... eg. we never say "if we have authentication, then..."
... true for all kind of constraints
... we said that we can reject a request
... since we don't have PATCH (only PUT), that's a real problem
... are people confused about what's been proposed?
... we need to make progress

bblfish: some much change going on in the docs, need time to read it
... the client need to know when there is an exception
... it's like non-monotonic things
... maybe put things in the headers?

Arnaud: john's email list all the sentences that would change
... I think that people just didn't take the time to review those

<bblfish> what is the url for John's e-mail?

Arnaud: we reached the LC, was exciting, but we're not done

<JohnArwe> http://lists.w3.org/Archives/Public/public-ldp-wg/2013Sep/0067.html

<JohnArwe> (also linked to from agenda)

<bblfish> thanks

Arnaud: please don't be silent, ask for clarifications

<stevebattle6> Bye

<BartvanLeeuwen> thx

<bblfish> bye

<Arnaud> thanks

<Arnaud> trackbot, end meeting

- DRAFT -

Linked Data Platform (LDP) Working Group Teleconference

14 Oct 2013

Attendees

Contents

Summary of Action Items

Scribe.perl diagnostic output