Re: Do we have consensus that we don't need more R2RML syntaxes?

David,

On 21 Jun 2011, at 23:45, David McNeil wrote:
> I am trying to think through this claim in relation to the charter text: "The mapping language SHOULD have a human-readable syntax as well as XML and RDF representations of the syntax".
> 
> I suppose the two can be reconciled if an implementation is required to support all of the syntaxes?

Which other syntaxes besides Turtle are you thinking of? I guess you're thinking about an XML-based syntax in particular? The requirements for human-readability (to some extent) and RDF are satisfied by Turtle.

> Another possibility is that is it not failure to have multiple, optional representations?

We could add a sentence:

[[
Conforming implementations MAY support other RDF serializations besides Turtle.
]]

Do you think it's worth stating this explicitly?

To be honest, I was thinking that this is obviously true even if left unstated, and I think it's obvious that many implementations will want to do this.

But adding the sentence wouldn't hurt.

> Or I suppose the charter could be wrong?

It says “SHOULD”, so if there's a good reason not to do it, then it's ok not to do it.

> I am curious if anyone is aware of how this issue has been addressed by other standards which use RDF as their data model.

I'm not actually aware of that many examples of complete W3C standards (rather than just vocabularies, where the question of syntax really doesn't matter) that are based on RDF. And the multitude of syntaxes that we have now is also a new phenomenon -- until quite recently, RDF/XML was the only W3C-recommended game in town. So for the standards I can think of right now (RSS 1.0, POWDER) the approach has been to prescribe the use of RDF/XML, along with a custom media type (application/something+xml) and some constraints on the allowed way of serializing the RDF as XML.

In D2RQ, we prescribed Turtle. I'm still happy with that choice. The only thing that could tempt me into preferring another syntax would probably be a custom syntax that smells like SPARQL.

Best,
Richard

Received on Tuesday, 21 June 2011 23:20:34 UTC