Turtle update (was Re: RDF Syntaxes 2.0)

Michael Schneider wrote:
> Hi Jeremy!
> 
> Jeremy Carroll wrote:
> 
>> On TTL
>>
>> I would have thought that if we decide to have another RDF WG, turtle
>> would be a candidate for an attempt on the W3C speed record from opening
>> the WG to Rec .... e.g. month1 FPWD, month 2 LC, month 3 PR.
> 
> Isn't this a little bit too optimistic? I think I remember some initial
> ideas about a very rapid OWL working group (of a year or so), backed by the
> fact that, at that time, there has already been a pretty complete and
> good-quality set of member submissions? But you know that things went
> differently.

I wouldn't be as optimistic as Jeremy, but we (well me) might produce
another team or member submission doc that might be closer.

> Specifically for Turtle, I do not really believe that everything will work
> out that easily, only because there already exists a team submission,
> because there has been some substantial criticism on the document, e.g.:
> 
>   <http://lists.w3.org/Archives/Public/semantic-web/2008Jan/0128.html>

That's the longest comment and I was aware of it.

It does address several flaws in the document, but implementors don't
seem to have had too much problem using the test cases to figure it out.

> I would believe that a future RDF WG would have to take comments like these
> into account, before being able to create a Last Call document. And, after
> the latest RDF thread in the SWIG list, I do not have the feeling that there
> is much consensus on any of the features that have been discussed. I rather
> think I have heard different opinions for most of them. So this looks to me
> as if a lot of discussion will have to take place in a WG, before one can
> even start drafting anything. Frankly, I think that, after one or two
> months, the WG will be happy if it has managed to settle on how to organize
> their issue list. ;-)

Formally, I don't think a WG has to take into account pre-WG document
comments, although it would be reasonable for the same comments to be made.

> Please don't get me wrong, I do not talk against bringing Turtle on Rec
> track here (actually, I do not have an opinion on this yet). I'm just having
> serious doubts about the time line you suggest.
> 
> Apart from this, it doesn't make much sense to me to have some RDF
> serialization in PR before having the abstract syntax / data model, on which
> it depends, in its final form. For example, the WG may discuss whether to
> allow for some form of generalized RDF (bNodes and/or literals in more or
> all positions of a triple) or to keep with all the syntactic restrictions of
> RDF 1. At least, Rec track serializations should follow this decision, i.e.
> they should neither allow for more general RDF (would be unjustified) nor
> for less (would prevent expressing all of RDF). 
> 
> So a W3C Recommendation for Turtle can hardly be published earlier than the
> core of the RDF spec. And by the "core spec" I refer to the RDF abstract
> syntax, the RDF vocabulary (the normative set of RDF(S) URIs such as
> "rdf:type"), and the RDF semantics.

I've consolidated the issues I've got from my inbox, comments and elsewhere
(IRC) into an issues list.  That github repo also has the turtle spec
history and tests.

http://github.com/dajobe/turtle/blob/master/ISSUES.md

(might improve the layout, I'm new to Markdown)

and then I might make a new version there if I have the energy. It's
more likely I'll do this before any new WG appears, I would guess.

Dave

Received on Tuesday, 26 January 2010 01:49:40 UTC