JSON-LD Community Group Telecon

Minutes for 2013-08-06

Agenda
http://lists.w3.org/Archives/Public/public-linked-json/2013Aug/0014.html
Topics
  1. Updates to Syntax Spec by David Booth
  2. GSoC update from Vikash
  3. Review JSON-LD github issues ready to be closed
  4. Review all LC2 and post-LC2 RDF WG issues
  5. Candidate Recommendation Preparation
Resolutions
  1. Interpret objects that do not have a @context entry as the JSON-LD Context when passed into the API functions (via any context parameter). When passing in an array of objects and strings, the same rule applies. Remote context documents specified via a URL are still required to contain an @context key to be a valid JSON-LD Context.
  2. RDF WG issue 129, 130, 132, 133, 134, and 135 have been addressed by the group and are resolved. Manu will send out official responses.
  3. The JSON-LD test suite will be a living test suite (updated as needed). The version of the test suite when we transition into Candidate Recommendation will be assigned a git tag, so that others can test and report against a static version of the test suite.
Chair
Manu Sporny
Scribe
Markus Lanthaler
Present
Markus Lanthaler, Manu Sporny, Vikash Agrawal, Dave Longley, Gregg Kellogg, Niklas Lindström, Paul Kuykendall, David I. Lehn
Audio Log
audio.ogg
Markus Lanthaler is scribing.
Manu Sporny: Today we'll re-review all issues and prep for Candidate Recommendation.

Topic: Updates to Syntax Spec by David Booth

Markus Lanthaler: These are the changes above ^^ [scribe assist by Manu Sporny]
Markus Lanthaler: Mostly editorial changes, and David Booth agreed to language offered by Dave Longley [scribe assist by Manu Sporny]
Manu Sporny: Any objections to pulling both those change request in? [scribe assist by Manu Sporny]
Manu Sporny: No objections from the group, ok, let's do that.

Topic: GSoC update from Vikash

Vikash Agrawal: hi Everyone, I received a wake up call email from Manu today and I do realise losing of time and effort and not getting as much done as I need to. I apologise for not able to write my weekly update email but I was drafting one last night. But after I read the e-mail, I think it makes more sense, if I write it next week. Also regarding the creator tool, I am able convert the details from form to JSON but using valid Context, how should convert to JSON-LD. Is there a library for this?
Vikash Agrawal: Also regarding the context, I have been naive but I am progressing. Apologies.
Manu Sporny: vikash, you don't convert from JSON to JSON-LD - you just add a context and the JSON /becomes/ JSON-LD. [scribe assist by Manu Sporny]
Markus Lanthaler: vikash, what library exactly are you looking for?
Manu Sporny: also, we're having trouble understanding what problem you're trying to solve. [scribe assist by Manu Sporny]
Markus Lanthaler: viksash, let's discuss this on the mailing list
Markus Lanthaler: it's pretty difficult to do so now on IRC
Manu Sporny: vikash - let's take the discussion offline, it's going to be difficult to talk through this via IRC while on the telecon. [scribe assist by Manu Sporny]
Vikash Agrawal: http://bit.ly/13D6iIZ
Vikash Agrawal: ok

Topic: Review JSON-LD github issues ready to be closed

Manu Sporny: markus, you said we support passing objects having a @context member in and did so from the very beginning. is that correct?
Markus Lanthaler: yes
Manu Sporny: and do implementations support that?
Markus Lanthaler: yes I think all of them do.. not all implementations support passing in arrays (mine for example doesn't; dave's does)
Dave Longley: right. my implementation looks for a @context key. if there's one, it is used, if not the object is used directly
Manu Sporny: markus, are you saying we shouldn't support that?
Markus Lanthaler: no.. we have to decide that. the danger is that a document is interpreted as a context without throwing an error
Gregg Kellogg: if I pass in an object to my context processing algorithm I presume it is a context not a context document
Gregg Kellogg: this implies that the normal flow is to pass objects having an @context member.. this is the exceptional case
... perhaps we should invoke a warning callback
... this is a case where I would like to see a warning
Dave Longley: this doesn't seem unusual to me
Niklas Lindström: the problem with allowing arrays of objects with @context is that it might process "invalid" documents as contexts
... I can't see a use case to support that
Manu Sporny: people may have a list of context URLs and the dereference them and replace them with the content of those documents
Gregg Kellogg: we can't test that without creating specific API tests
PROPOSAL: Interpret objects that do not have a @context entry as the JSON-LD Context when passed into the API functions (via any context parameter). When passing in an array of objects and strings, the same rule applies. Remote context documents specified via a URL are still required to contain an @context key to be a valid JSON-LD Context.
Manu Sporny: +1
Dave Longley: +1
Niklas Lindström: +1
Gregg Kellogg: +1
Paul Kuykendall: +1
David I. Lehn: +1
Markus Lanthaler: +0.1 (don't see much value in specifying this and it adds further variability)
RESOLUTION: Interpret objects that do not have a @context entry as the JSON-LD Context when passed into the API functions (via any context parameter). When passing in an array of objects and strings, the same rule applies. Remote context documents specified via a URL are still required to contain an @context key to be a valid JSON-LD Context.
Manu Sporny: I think we're done w/ this. [scribe assist by Manu Sporny]
Markus Lanthaler: ok, I'll close 254. the official response will be handled using the issue in the RDF WG tracker
Manu Sporny: I've updated an example in response to Robin's feedback
... I hope it clarifies it
Dave Longley: I've read the whole discussion and agree
Manu Sporny: the last major (editorial) change to make is to briefly describe JSON-LD's data model at the beginning of the spec
... I'll sit down and re-arrange pieces and the spec (not removing changes Peter-Patel S. and David B. made) to make the spec easier to read
... I'll introduce the concepts at the high level at the beginning of the spec
... I don't think there's much to discuss
... I'll make the changes and we can discuss them afterwards
... if there are no objections I'll go ahead and make those changes
... that's the only issue remaining for the syntax spec. Markus, all issues for the API spec have been addressed, right?
Markus Lanthaler: yes, only the @context-change we'll need some minor tweaks in the API spec

Topic: Review all LC2 and post-LC2 RDF WG issues

Markus Lanthaler: sandro already outlined the steps we should take: http://lists.w3.org/Archives/Public/public-rdf-wg/2013May/0260.html
Manu Sporny: I don't necessarily agree but OK
... but it won't prevent us from going to CR and that's the main point at the moment
... ISSUE-132: both Peter and David said they are happy with the changes we made
... ISSUE-133: reverse properties are a feature at risk
... ISSUE-134: blank node graph names are now supported in RDF; we discuss blank node properties in the spec
... ISSUE-135: we addressed all the feedback. We still need to send the official response
PROPOSAL: RDF WG issue 129, 130, 132, 133, 134, and 135 have been addressed by the group and are resolved. Manu will send out official responses.
Paul Kuykendall: +1
Manu Sporny: +1
Dave Longley: +1
Markus Lanthaler: +1
Niklas Lindström: +1
David I. Lehn: +1
RESOLUTION: RDF WG issue 129, 130, 132, 133, 134, and 135 have been addressed by the group and are resolved. Manu will send out official responses.

Topic: Candidate Recommendation Preparation

Manu Sporny: I think we are ready to go to CR
... we have no other comments to address
... the only thing we'll have to do is to prepare the docs
... I'll take care of the syntax spec. Markus can you prepare the syntax spec
Markus Lanthaler: sure.. what date do we tag the documents with?
Manu Sporny: August 20
Markus Lanthaler: there's no RDF WG telecon next week
... the next meeting is August 21
Manu Sporny: ok.. let's try to publish on August 22 then
Paul Kuykendall: can we send an implementation report for a propertary implementation?
Manu Sporny: sure
Paul Kuykendall: do we see any other major or minor changes?
Manu Sporny: no, we would need to go through another last call then. we do not expect many changes
Markus Lanthaler: pkuyken, here are links that might help you: http://json-ld.org/test-suite/ and http://json-ld.org/test-suite/reports/
Gregg Kellogg: during CR we may also want to issue a call for implementations
Manu Sporny: should we discuss this now or during CR?
Gregg Kellogg: I suggest we keep the test suite where it currently is and update it even post publication
Manu Sporny: we can freeze the test suite at any point in time by using the commit hash
Gregg Kellogg: so which URLs should we use in the implementation reports? A URL including the git hash?
Manu Sporny: yes.. I think that would make sense
... I think we did the right thing with the RDFa test suite which is updated as soon as an issue is found
PROPOSAL: The JSON-LD test suite will be a living test suite (updated as needed). The version of the test suite when we transition into Candidate Recommendation will be assigned a git tag, so that others can test against a static version of the test suite (if required).
Paul Kuykendall: +1
Dave Longley: +1
Gregg Kellogg: +1
Niklas Lindström: +1
Manu Sporny: +1
Markus Lanthaler: +1
David I. Lehn: +1
RESOLUTION: The JSON-LD test suite will be a living test suite (updated as needed). The version of the test suite when we transition into Candidate Recommendation will be assigned a git tag, so that others can test and report against a static version of the test suite.
Markus Lanthaler: do we need to add the exit criteria to the specs?
Manu Sporny: yes.. I can take care of that or you just look at other specs in CR.. basically we just need to say that we require two implementations pass every test and an implementation report
Markus Lanthaler: ok.. how long will we stay in CR?
Manu Sporny: I think we agreed on 4 weeks