Re: simplerdf & Towards the future RDF library

On Sat, Sep 12, 2015 at 2:43 AM, Ruben Verborgh <ruben.verborgh@ugent.be>
wrote:

> Hi,
>
> > can we all try to hang out on #rdfjs irc.w3.org or
> https://gitter.im/rdfjs ?
> > maybe we can try scheduling a telecon in next weeks?
>

IRC is preferable. For conference calls with Zakim, it seems to be required.


>
> That would be awesome. Let's try to gather all involved people.
> We'd probably need to prepare this in a structured way though,
> as this is a big effort obviously.
>
> We'd need to build a charter, including:
> – desired features
> – desired supported languages
> – architectural focus
> – target platforms
> – practical agreements (code style, commit/branch style)
>
> We could use this group's infrastructure for that.
>

Yes, please! This would probably best be suited for a teleconference where
we can prioritize.

Personally, I'm interested in resuming work on RDF Interfaces, which is
certainly under the scope of this CG if nothing else. We would continue the
WebIDL definition, targeting ECMAScript, but remain platform-neutral, the
same way how DOM has implementations in Java, PHP, etc.

Some sort of asynchronous interface would be nice, too, especially if
combined with a higher-level query interface for remote servers (like
SPARQL servers). I also defined an "IndexedGraph" idea (think IndexedDB),
see <https://lists.w3.org/Archives/Public/public-rdfjs/2013Oct/0040.html>.

Specifications should be accompanied by test cases. There's now an RDF
Tests Community Group, and a repository at <https://github.com/w3c/rdf-tests>.
I imagine we would define many of our own tests, and also pull tests from
wherever it's appropriate.

I've implemented RDF Interfaces at <https://www.npmjs.com/package/rdf> and
also a library for IRI parsing and manipulation, <
https://www.npmjs.com/package/iri>. They're in need of modernizing, though.

I'm also working on JSON Schema at the moment, which is all hypermedia. In
one of my applications, I parse JSON documents against their schema,
extract link relations, and store it with the 'rdf' package. This is the
sort of interoperability I'd like to advance.

Austin.

Received on Sunday, 13 September 2015 21:31:49 UTC