RE: Reducing the number of keywords (ISSUE-15, ISSUE-26, ISSUE-31)

> >> I'd also favor a "@suffix" element, so that I could use
> >>
> >> {
> >>     "@base": ""
> >>     "@suffix": "#id"
> >>     "friend":  {
> >>             "id": "foobar",
> >>     }
> >> }
> >
> > I can't see what kind of functionality a @suffix element would
> allow.. Could
> > you elaborate?
> >
> That would let JSON authors remove #whatever fro JSON-LD values - and
> have those automatically added when transforming JSON-LD to RDF

That would add it to every IRI. Honestly, I don't see much value in it. Also
it makes understanding to which IRI "foobar" is eventually expanded quite
difficult as you would have to remember even more things. That being said we
assume that the vast majority of JSON-LD documents will be generated by
machines and not written manually by humans. So it is much more important to
make them easy to read then to write. Being more explicit and reducing the
number of things to keep in mind surely helps in that regards.

> > Same here.. what kind of functionality does this enable? I can see
> some use
> > cases where URI templates are needed but that can be described
> without
> > having explicit support by JSON-LD as well IMO.
> Same as before - is there already some templating support in native
> JSON ?

No.. But you could define a concept whose value is a template. Or even a
datatype for templated IRIs.

> Overall my goal is to make the JSON serialisation as simple as
> possible - while keeping the RDF-ization of JSON-LD documents as
> complete as the spec allows.

So is mine. In the last telecon we made great progress in greatly
simplifying JSON-LD while still keeping all of the functionality. As far as
I know, we currently support everything that's supported in RDF.

Markus Lanthaler

Received on Thursday, 15 December 2011 15:57:08 UTC