Re: JSON-LD Telecon Minutes for 2014-01-07

Sorry, I guess I dropped the ball on organizing a meeting for today. Sounds like you all had a useful discussion, though. I'd say that, in the future, we continue discssion on the mailing list/IRC and when we agree we need to get into it more, we'll arrange a telecon.

Gregg Kellogg

On Jan 7, 2014, at 8:27 AM, Markus Lanthaler <> wrote:

> Despite not sending out a formal agenda we had a brief telecon today. 
> The minutes are now available.
> A full transcript of the meeting can be found below.
> The audio will be added soon.
> -------------------------------------------------------------------
> JSON-LD Community Group Telecon Minutes for 2014-01-07
> Agenda:
>   n/a (ad-hoc telecon)
> Topics:
>   1. Tooling for JSON-LD
>   2. Context at
>   3. Processing of relative IRIs without base
>   4. Subtree split to create a repository containing just the 
>      JSON-LD tests
> Chair:
>   Markus Lanthaler
> Scribe:
>   Markus Lanthaler
> Present:
>   Niklas Lindström, Markus Lanthaler, Paul Kuykendall, Dave Longley
> Audio:
> Niklas Lindström: .. Is there a telecon today? We planned for the 
>   7th last time, but there's been no mail about it.
> Markus Lanthaler: I don't know myself.. I just joined but was the 
>   only one
> Paul Kuykendall: I was wondering about that myself
> Markus Lanthaler: Do we have something to talk about?
> Paul Kuykendall: We discussed last time the "next steps" with 
>   respect to tooling, etc.
> Paul Kuykendall: Do we have any other items that we should go 
>   over?
> Niklas Lindström: I guess we ended by saying something like "of 
>   course, unless we have a bunch of topics/issues, we could wait 
>   until we've gathered some and then issue a call"..
> Markus Lanthaler: There was some discussion regarding the context 
>   at but other than that I don't think anything else 
>   happened
> Markus Lanthaler:  I agree, I think that's about it right now. 
>   [scribe assist by Niklas Lindström]
> Markus Lanthaler: So I guess we postpone the telecon!?
> Paul Kuykendall: Has anyone/group come up with a wish-list of 
>   tools, etc. that would be beneficial?
> Niklas Lindström: I've no problems with taking that on the list 
>   and see if we can coalesce the questions around that (publishing 
>   contexts, caching and other tooling)
> Markus Lanthaler: pkuyken.. do you want to briefly discuss this 
>   on the call? I'm happy to have a brief call but also fine with 
>   moving it to the mailing list
> Paul Kuykendall: dialing in
> Markus Lanthaler: niklkasl, will you join us as well for a couple 
>   of minutes?
> Niklas Lindström: sure, dialing in
> Markus Lanthaler: ping m4nu, taaz, dlongley_
> Dave Longley:  we are having a brief call if you want to join 
>   [scribe assist by Markus Lanthaler]
> Dave Longley: i can join for a bit
> Markus Lanthaler is scribing.
> Topic: Tooling for JSON-LD
> Paul Kuykendall: 2) discussion on list
> Paul Kuykendall:  during the last telecon we discussed what 
>   tooling we wanted to have
>   ... I'm wondering where we are at
>   ... do we want to put something up on or on the 
>   wiki on GitHub?
> Niklas Lindström: is having sound issues, so I may miss some 
>   speech from time to time
> Markus Lanthaler:  me too
> Dave Longley:  I think it would be good to have a primer
>   ... David Lehn was working on it some time ago but it's not 
>   done yet
>   ... people have issues understanding some things (like 
>   overloading of @type)
>   ... not sure if a primer would help
>   ... regarding tools: I think creating a wiki page and linking 
>   it from would be a good start
> Niklas Lindström: +1 for a wish list (and list of common 
>   questions)
>   ... also updates for the playground would be nice
> Paul Kuykendall:  a simple copy button on the playground would be 
>   awesome
>   ... where shall we track these things?
> Dave Longley:  regarding the playground I think you should just 
>   file an issue
> Paul Kuykendall:  the only concern I have with the wiki (w3c 
>   wiki) is that people have the feeling to have to dig into specs
>   ... instead of just having to use tools etc.
> Dave Longley:  I think we could just use Github issues for this 
>   as well
> Niklas Lindström:  one difficult question is always if JSON-LD is 
>   usable as just JSON
>   ... might be an example for that.. can people use 
>   different terms (if they are properly mapped in the context)?
> Dave Longley:  I think we should focus on the JSON side of things
> Niklas Lindström:  yeah.. we should mention what kind of 
>   constraints that imposes in the primer
> Dave Longley:  do people agree with paul that we should avoid the 
>   wiki
> Niklas Lindström:  yeah, wikis are lousy for discussion
> Markus Lanthaler:  I agree. I would prefer to just use Gihub 
>   issues with a specific tag so that we can directly link to that 
>   list
> Dave Longley:
> (etherpad could be better but issue is good and available)
> Markus Lanthaler:  I think we shouldn't format it using ReSpec 
>   but make it look more like a blog post
> Dave Longley:  yeah
> Topic: Context at
> Paul Kuykendall:  do we want to discuss the context at 
>   now?
> Markus Lanthaler:  I don't think we can discuss much here because 
>   (unfortunately) we are not in control there
> Niklas Lindström:  one of the interesting questions to me is that 
>   at this stage the intent of the examples being published is to 
>   create data being published by Google
>   ... I think this intended for publishers to publish data for 
>   Google and the other search engines.. not other people
>   ... Google obviously will have their own caches
> Markus Lanthaler:  most problems could be addressed by using 
>   @vocab
> Dave Longley:  that wouldn't address the problem that Martin Hepp 
>   is throwing in (tools accessing various URLs for 
>   properties etc.)
> Niklas Lindström:  that might be a problem because tools might do 
>   optimizations like prefetching etc.
>   ... there are no such tools available for JSON-LD (yet)
> Markus Lanthaler:  that's a general Linked Data "problem"
>   ... tools like Tabulator etc. will have to dereference 
>   properties etc. to get their labels
> Dave Longley:  in a lot of cases that's the whole point of all 
>   this
>   ... anyway. I think there's progress been made on the mailing 
>   list
> Topic: Processing of relative IRIs without base
> Dave Longley:  markus, I think we never decided what happens when 
>   base is set to null
> Markus Lanthaler:
> Markus Lanthaler:  at least my understanding from reading the 
>   RFCs is that the algorithms can only be run if there's a base
>   ... if @base is set to null then there's none
> Dave Longley:  is there not even something like path 
>   normalization we might could run
> Markus Lanthaler:  I don't think so.. the algorithms are not 
>   defined
>   ...
>   ... The term "relative" implies that a "base URI" exists 
>   against which the relative reference is applied. Aside from 
>   fragment-only references (Section 4.4), relative references are 
>   only usable when a base URI is known.
>   ... A base URI must be established by the parser prior to 
>   parsing URI references that might be relative.
> Dave Longley:  I'm fine with not touching relativ IRIs if there's 
>   no base
> Markus Lanthaler:  it's a bit weird but should we add a test for 
>   this?
> Dave Longley:  hmm.. no I don't think so
> Markus Lanthaler:  it's weird but we don't need to use an empty 
>   string
> Dave Longley:  yeah.. maybe that makes sense
> Markus Lanthaler:  that brings me to something else...
> Topic: Subtree split to create a repository containing just the JSON-LD
> tests
> Markus Lanthaler:   could we create a subtree split including 
>   just the tests?
> Paul Kuykendall:  I think that would be a great idea
> Markus Lanthaler:  do you have access to set up a post-commit 
>   webhook on GitHub? I don't
> Dave Longley:  I have access
>   ... I'll try to give you access as well
> --
> Markus Lanthaler
> @markuslanthaler

Received on Tuesday, 7 January 2014 18:11:26 UTC