Re: RDF Interface specification

Hi all,


On Thu, Jul 3, 2014 at 1:49 PM, Melvin Carvalho <melvincarvalho@gmail.com>
wrote:

>
>
>
> On 2 July 2014 16:36, Adrian Gschwend <ml-ktk@netlabs.org> wrote:
>
>> hi group,
>>
>> For about two years the RDF Interfaces spec is in limbo:
>>
>> http://www.w3.org/TR/rdf-interfaces/
>>
>> It states:
>>
>> "This document is not finished. Due to the lack of time, the RDF Web
>> Applications Working Group was unable to complete work on this document
>> before the end of their charter. At the time of publication of this
>> document, it was not known whether W3C will continue this work on the
>> Recommendation track in another Working Group. While a significant
>> amount of design work went into this document, at present there are no
>> known implementations of the specification. A number of design issues
>> have not been completely resolved. Developers wishing to implement this
>> API should be aware of incomplete nature of the specification."
>>
>> However, in the JavaScript world RDF Interfaces is by now the base for
>> quite some projects and gets ongoing development by multiple developers.
>> There are at least 3 implementations available.
>>
>> Last week Thomas Bergwinkl announced some new JavaScript libraries and
>> applications which build heavily on top of RDF Interfaces.
>>
>> Announcement:
>>
>> http://lists.w3.org/Archives/Public/public-rdfjs/2014Jun/0008.html
>>
>> Followup with motivations:
>>
>> http://lists.w3.org/Archives/Public/public-rdfjs/2014Jul/0001.html
>>
>> He also proposes a new specification called RDF-Ext [1][2], which
>> extends RDF Interfaces with a store interface, async parsers &
>> serializers and an ES6 Promises interface.
>>
>> The question now is what would be the best way to work on those
>> specifications. Currently it is published on a github page and the
>> content is residing in a github repo.
>>
>> Is it the idea of W3C to standardize specifications which are more or
>> less language dependent? While RDF Interface never clearly stated it, it
>> is heavily oriented on JavaScript syntax and for that reason it most
>> probably won't be implemented in other languages. The same is true for
>> RDF-Ext. However, Thomas and I are interested in making it "official",
>> in case the group thinks that this makes sense.
>>
>> I initiated the RDF JavaScript Community Group [3] for that reason so we
>> would also offer to take over both RDF Interfaces and RDF-Ext
>> specification development, if this makes sense.
>>
>> comments would be appreciated
>>
>
> Glad to see ongoing work on this project!
>
> The RWW Community Group has already agreed to incubate this spec 1-2 years
> ago.
>
> Adrian, Thomas and Nathan are already participants in this group. [1]
>
> My understanding was that this spec was essentially complete.  From what I
> remember it was on REC track, pending some minor details to RDF 1.1.
>
> I think Nathan was a bit pushed for time to support and maintain this
> work, and from what I remember it was changed from REC track, to Note, due
> to what was felt to be insufficient resources ...
>
> Now that RDF 1.1 is baked, it should be possible to finish this work, if
> indeed it's not finished already.
>

It is correct that it was not finished. There were a number of open issues,
regarding naming of interfaces and properties in relation to RDF concepts,
and possibly complexity of use (though that was at least in part due to the
related notes for RDF and RDFa APIs ([1], [2]), which where even less
complete).

There were fears that the releases of these as Notes, even with the clear
warning stating that is is work-in-progress that should not be relied upon
for interoperability, would still be interpreted as a message that it
represented consensus in design.

I support Adrian's suggestion that the RDF-JS Community Group is the
appropriate place to pick up this spec (and related work). We should do so
fully knowing that it is not finished, and be open to critique regarding
naming, effectiveness and complexity. We have more wild usage of RDF in JS
now, and more people thinking about these matters in general. So I think
the time is ripe to start discussing what we need to have standardized.
(And what to leave up to community libraries to explore – to allow patterns
to emerge and mature. Premature standardization can certainly stifle a
technology. We need to measure this with real use cases and data in real
projects.)

Considering the relationship of this with JSON-LD – both the syntax [3] and
the API [4] – is important, since they are already RECs, and it would be
wasteful to use diverging mechanisms for similar problems. (We also begun
looking into an idea for connecting raw JSON-LD objects into a graph
structure [5], though it wasn't finished.)

Personally I found the use of expanded JSON-LD as raw, interoperable
representation between toolchains in Javascript to be rather lightweight.
But that was some time ago, and alas I haven't spent much time lately to
assess the effectiveness of these interfaces in the various new libraries
that have been created since.

(I also have some old experiments ([6], [7]) based on this, mostly oriented
towards revising the RDF API design towards a JSON-LD context-based
mechanism. For real projects lately, I have mostly used compacted JSON-LD
within JS to get effective shapes and simplify code. In these scenarios, I
use a lightweight version of these experiments to get the proper graph
shape in place. Of course, there are many different ways to layer these
designs.)

Cheers,
Niklas

PS. Should public-rdfjs@w3.org be included in this thread as well? I'll
forward this message for now.

[1]: http://www.w3.org/TR/rdf-api/
[2]: http://www.w3.org/TR/rdfa-api/
[3]: http://www.w3.org/TR/json-ld/
[4]: http://www.w3.org/TR/json-ld-api/
[5]: http://json-ld.org/spec/latest/json-ld-connect/
[6]: http://niklasl.github.io/rdfa-lab/examples/rdfa-api-comparisons.html
[7]: https://github.com/niklasl/rdfa-lab/wiki/RDFa-DOM-API



>
> The RWW would be really happy to help out any way we can.  Alternatively
> another community or working group could pick up the ball, if wished.  Very
> happy to help here, just let us know what you feel need doing or the next
> steps
>




>
> [1] http://www.w3.org/community/rww/participants
>
>
>>
>> regards
>>
>> Thomas & Adrian
>>
>>
>> [1]: https://bergos.github.io/rdf-ext-spec/
>> [2]: Implementation of it: https://github.com/bergos/rdf-ext
>> [3]: http://www.w3.org/community/rdfjs/
>>
>>
>

Received on Thursday, 3 July 2014 12:42:18 UTC