Re: social-ISSUE-45 (mf2jsonldconflicts): Conflicts between json-ld and mf2 examples [Activity Streams 2.0]

On 11 August 2015 at 00:42, Sarven Capadisli <info@csarven.ca> wrote:

> On 2015-08-10 20:37, Social Web Working Group Issue Tracker wrote:
>
>> social-ISSUE-45 (mf2jsonldconflicts): Conflicts between json-ld and mf2
>> examples [Activity Streams 2.0]
>>
>> http://www.w3.org/Social/track/issues/45
>>
>> Raised by: Benjamin Roberts
>> On product: Activity Streams 2.0
>>
>> As was mentioned by https://www.w3.org/Social/track/issues/44 there is a
>> disparity between several examples in AS2 between JSON-LD and MF2.
>>
>> Examples should be reviewed on a case-by-case basis considering
>> A) Can AS2 be improved to better represent pragmatic social web
>> publishing and consuming experience?
>> B) Are there real world use-cases implied by AS2 features that need
>> documentation as input for proposing new microformats2 properties (or
>> possibly objects)
>> C) A combination of both A and B
>>
>
> ISSUE-44 is about the removal of the microformats (mf) examples from the
> AS2, because 1) ACTION-26 is incomplete, and there is no good sign for it
> to be complete with reasonable quality as part of a W3C Recommendation
> because, 2) it is out of place; i) the mf examples do not demonstrate a
> "generally equivalent" "serialization" of the JSON-LD examples, ii) do not
> demonstrate the semantics underlying AS examples, and iii) there is no work
> in the mf wiki which can reasonably demonstrate the AS2's breadth and depth
> coverage of "social" activities. This is still acknowledging that AS2 is a
> work in progress.
>
> Meanwhile, this issue's (ISSUE-45) core concern is to improve the AS2
> examples, in order to improve the mf examples. That proposal is flawed and
> backwards. The example statements/activities are what we start with, i.e.,
> the human language describing the desired social Web activities (e.g.,
> based on interest, existing practices on the Web, business cases, or
> whatever else there may be). Producing a machine-processable implementation
> is the second step. What you are proposing is fundamentally to see how can
> we reshape the examples in order to fulfil the limitations of the code
> based on an external community's (mf) centralized vocabulary development
> and process. If how we represent those examples in code is critical, then
> the simplest solution (and the one in which requires the least amount of
> effort) is to re-use what's available from the existing W3C
> Recommendations. One simple consideration to make here is to see whether
> the toolbox you want to work with is appropriate.
>
> Specific to your proposals points:
>
> A) Define "pragmatic social web publishing and consuming experience", and
> explain why that is the criteria that you have selected, and well as what
> other criteria that you have considered, and why they were dismissed. If
> you can address these questions, then we can better understand what you
> mean by your proposal to improve the current AS2.
>
> As per your proposal to improve, which parts of AS2 have you implemented?
> What are your limitations with AS2? Can you create GitHub issues for them?
>
> B) Start with what is already in the current AS2 examples. Take it to the
> mf community, run it through the mf process. Bring back whatever passes the
> mf process, place them in respective AS2 MF examples.
>

This was discussed during the call, unfortunately we were quite pushed for
time.  Lots of good points were raised, which I dont think were completely
captured in this conversation.

Arnaud's major objection to issue 44 seemed to be with the formulation of
the title.  e.g. that it was phrased more as a proposal than an issue.
Seemed to be a pretty valid point.

The suggestion was to merge 44 + 45 together, however, it was pointed out
that 44 and 45 are quite unrelated.

Another suggestion was to reformulate the title into the issue that AS2 and
MF2 are not aligned.

The other point made was that the examples are informative and not
normative.

To this extend I could imagine a stance of "this is something I dont
particularly like, but could live with" happening.  My concern is that work
on improving MF2 slows down AS2 from getting to CR.

I think the misalignment of AS2 and MF2 is one part of a bigger question,
and that is whether MF2 as a vocab should have special treatment in the
spec, largely due to the WG having a lot of MF2 voting members.

Things I'd like to see I see with MF2 in general are:
- A stable reference vocab document, much like the AS2 vocab
- Alignment between AS2 and MF2 terms
- Clear understanding of how the MF2 process would work
- Machine readable interpretations of the terms, as per AS2
- Alignment with existing W3C RECs (ie linked data) as per AS2

But even if all these were achieved, should MF2 get special treatment,
over, say : schema.org, facebook open graph (which have more adoption), or
FOAF/SIOC.  Probably still not, imho.  I can see it being valuable in a
primer or not, but in the REC itself, would it just be confusing?  Hard to
say.

All these things seem to be a "nice to have", rather than, a must.  Would
be nice to get AS2 to CR and then look towards implementations and test
suite.

My take on this is that I dont think MF2 should be in the doc, but it's
something I could live with.  I'd like to see improvements, but not to the
extent that AS2 gets slowed down by having an MF2 dependency.


>
>
>
> PS: There is now (today) a proposal to publish Activity Streams 2.0 as a
> Candidate Recommendation.
>
> -Sarven
> http://csarven.ca/#i
>
>