RE: RDF 1.1 Primer

On Wednesday, December 04, 2013 7:37 PM, Pierre-Antoine Champin wrote:
> On Wed, Dec 4, 2013 at 7:26 PM, Markus Lanthaler wrote:
> > I think it would be much more inviting if you would use example 11 as
> > first example in section C.2 JSON-LD, then show the context and mention
> > that it can also be embedded directly in the document and that prefixes
> > can be used the same way as in Turtle and use the multiple-graphs
> > example to illustrate it. I've cleaned up the examples and the context
> > and attached all of them in a single file.
>
> I thouht the same at first, but then I realized that the first example
> might be easier to understand *when compared to the others*. As I see it,
> the first example is using JSON-LD as "yet another" concrete syntax for
> RDF, and uses a straightforward way to encode a given graph. The second
> example assumes that the JSON was already existing in some API, and was
> RDF-ized with a JSON-LD context.

Hmmm... fair enough. In any case I think we need to decide who we are trying
to address with this example.

I guess most Semantic Web experts won't read the primer anyway and if they
do, they will be "patient" enough to wait for the second example. On the
other hand, for newcomers a representation similar to what they are familiar
with from JSON APIs may be much more appealing. At least they will
immediately see that it is possible to serialize RDF in such a way. I fear
that if we don't do this early on we will lose a lot of people.

I would even go as far as suggesting we show that directly in section 5.
Would anyone object to this?

 
> > > One quick question - any
> > > reason for mapping uri to @id instead of just reusing @id?
> >
> > As Pierre-Antoine already said it has advantages when working with the
> > data but I nevertheless think we probably should stick to @id and @type
> > in the primer as these are really advanced features.
>
> As explained above, the idea was to make it look as if it came from any
> API, not necessarily RDF- or JSONLD-aware But I won't fight over it if
> there is a consensus to switch back to @id

Fully agree with this but I think we should also keep in mind how most
examples look in the wild. Schema.org, Actions in GMail, json-ld.org, the
JSON-LD spec to just name a few all use @id and @type. If we alias them in
the primer, people will probably get confused when comparing it to the other
documents. I consider keyword-aliasing a really advanced feature.


> (you would probably want to switch to @type as well, then?)

Yes


--
Markus Lanthaler
@markuslanthaler

Received on Wednesday, 4 December 2013 18:56:37 UTC