toRdf and fromRdf algorithms need WebIDL API methods defined

[gkellogg]

[azaroth42]

Is this editorial, or does it need discussion?

[gkellogg]

Editorial, but we may discuss things in the PR review.

[BigBlueHat]

Mostly as an aside: the RDF.js CG is bringing in lots of code and working it toward conformance with the Interface Specification: RDF Representation. There's no existing conflict (that I've seen) between this WebIDL and their JS API, but it might serve both communities well if we keep each other appraised of any overlap (if/when it occurs).

Thanks for your work on this, @gkellogg!

[gkellogg]

@BigBlueHat thanks for the reference, I was aware of this, but had forgotten the details.

Technically, we should probably be using interfaces like RDF source, which the RDF Representation describes using the Source interface.

The Stream interface would seem to be better modeled using the WebIDL iterable construct, which is the primary thing we depend on for Convert to JSON-LD from RDF.

Interestingly, they don't really have the equivalent of an add method, which might be done in Sink, or through assigning to Quad.graph. However, they could consider adding it to the Sink interface.

But, the RDF Representation spec, I think, is more designed to describe actual JavaScript implementations, rather than an abstract model for dealing with RDF concepts.

[BigBlueHat]

But, the RDF Representation spec, I think, is more designed to describe actual JavaScript implementations, rather than an abstract model for dealing with RDF concepts.

Quite likely. The spec itself has also set dormant for quite awhile. I believe folks in the CG are hoping to revisit that work once they've consolidated some of the implementing code projects into the CG's org. But given that it will only ever (by design) be JS-centric, you're probably quite right. Something to keep an eye on (and people to work with), but probably not a reason to change course per se.