Re: AS2.0: JSON and/or RDF based?

Hi Harry, all,

On Sun, Feb 8, 2015 at 8:10 AM, Harry Halpin <hhalpin@w3.org> wrote:

> On 02/06/2015 11:58 PM, ☮ elf Pavlik ☮ wrote:
>
> > Personally I don't see conflict between JSON and RDF mostly thanks
> > to availability of JSON-LD. Currently AS2.0 not only uses JSON-LD
> > but AS2.0 Vocabulary also takes advantage of RDFS and even OWL.
>
> I think the charter is clear: It's JSON-based. Any use of RDF(S) or
> OWL inference is fine or alternative serializations is fine, but
> should not be required (and thus non-normative).


There's a difference between "not required" and non-normative: a feature
can be normatively not required by way of SHOULD or MAY (as you no doubt
realize).

A decision not to specify JSON-LD normatively at all will prevent it from
being used in linked data environments, as it would lack the JSON-LD
context that maps from JSON into RDF.  This would mean that the spec is
less likely to be used by other JSON-LD oriented specifications and
systems, such as in the Annotation WG, for the sake of not adding a single
key and value to the top level JSON object.

I don't think you can sit on the fence for this one, I'm afraid.

Rob




> If people want to
> document how Link Relations etc. can be aligned with RDF, that's fine
> and can be done informatively (as for some, it will definitely be
> useful) as long as it does not force consuming or producing
> applications to adopt a whole other data model.
>
>    cheers,
>       harry
>
> >
> > I see it similar to adopting aspects of well established Link
> > Relations, in our API work for example. It will not conflict with
> > charter JSON requirement as far as I can tell. Further aligning
> > Link Relations and RDF[2] can make it all even more straight
> > forward! Also LDP and Hydra, both mentioned as API candidates,
> > strongly incorporate both Linked Data and Link Relations. Linked
> > Data Fragments as well provide some solid *hypermedia* REST read
> > access interface. Last but not least - our charter also mentions
> > JSON-LD as potential data transfer syntax (-*LD*), right after
> > "describe the data using URIs in an extensible manner"...
> >
> > JSON provides well adopted serialization but we still need to work
> > with conceptual models. While many efforts go into experiments with
> > novel approaches. I see both W3C and IndieWeb having strong
> > tradition on building upon existing and established work.
>
>
-- 
Rob Sanderson
Information Standards Advocate
Digital Library Systems and Services
Stanford, CA 94305