Re: representing deserialized RDF - N3, JSON-LD, RDFa, XML/RDF

On 11 December 2013 00:37, Gregg Kellogg <gregg@greggkellogg.net> wrote:

> On Dec 10, 2013, at 3:06 PM, ☮ elf Pavlik ☮ <perpetual-tripper@wwelves.org>
> wrote:
>
> > Hi,
> >
> > I didn't yet look into all the parsers/serializers listed at our wiki,
> only in some parts of jsonld.js so far. I wonder if it would make sense to
> use common way of representing deserialized RDF in javascript. Ruben
> already touched topic of "URIs and literals" and I would like to take it
> all the way into representing datasets with named graphs, blank nodes etc.
> >
> > Also looking at excellent http://ruby-rdf.github.io makes me thinking
> that we could take little time in coordinating our efforts to provide more
> uniform toolbox. And luckily I believe we can count here on Gregg's
> experience on RDF.rb environment :)
>
> Of course.
>
> > Ruben works on N3 + Turtle, Alex on RDFa + Turtle, no one from people
> maintaining JSON-LD (js) in this group yet but we can still connect better.
> Not sure if anyone here works on XML/RDF
>
> Do we _really_ need RDF/XML in Javascript? It would be nice to move on
> with the more modern variations.
>

rdflib.js already has an rdf/xml parser

https://github.com/linkeddata/rdflib.js/blob/master/rdfparser.js

I suggest just having a common set of parsers that we can share

Having a quad store is important.

Test suites are also a plus and we have a few we can contribute.

CC: Joe Presbrey


>
> > I would find it very cool if while using all those parsers/serializes we
> could keep common representation of deserialized dataset. Also we could
> support each other when it comes to working with modeling some complex
> topics like: http://www.w3.org/TR/rdf11-concepts/#section-generalized-rdf|
> http://manu.sporny.org/2013/rdf-identifiers/ etc.
>
> In RDF.rb, we keep have a number of different classes and mixins to
> represent RDF information. The basic breakdown is the following structure:
>
> * Value, with subclasses Term, Graph, List and Statement.
> * Term, with subclasses Literal and Resource
> * Literal, with various sub-classes for different typed literals
> * Resource, with subclasses IRI and BNode
>
> There are also a variety of mixins to give Enumerable, Queryable, Mutable
> and so-forth behavior.
>
> In Javascript, I think if there was a common Statement class containing
> slots for subject, predicate, object and name, which could be a IRI, BNode
> or Literal, the various serializers and deserializers could generate or
> consume sets of these, which would provide one level of a common
> representation between the different tools.
>
> Gregg
>
> > Cheers!
> >
> > [1] http://www.w3.org/community/rdfjs/wiki/Main_Page
> >
>
>
>