Re: RDF 2 Wishlist: Turtle Syntax

2009/11/2 Axel Polleres <axel.polleres@deri.org>:
> ok, so this is a wishlist, so I am allowed to just add my "in an ideal
> world"-personal-my-private-little-hat-on favorites :-)
>
> I have the following:
>
> - Turlte has been mentioned exhaustively already... I'd also like to have a
> "Terse RIF syntax": I think a more RDF based
>  syntax for RIF is still missing for RIF being picked up by the RDF
> community.

+1
I'm fairly sure I'm not the only person around here that gets a bit
scared by RIF's presentation - an easy syntax would make it a lot more
accessible.

 If that meant rubbber-stamping N3 and
>  nailing down its corresponding RIF fragment, that could be a good starting
> point.

I suspect rubber-stamping N3 might be a bit of a rathole, though if
anyone has the confidence to take it on, it would be lovely to see.


>
> - Annotations for RDF: apart from quads/named graphs, you want to give
> arbitrary annotations to RDF triples
>  (time, provenance, trust, etc.). Is reification (without any real
> reification semantics) which is the
>  only existing way to support annotations, really the right way to solve
> this?
>  I think that would be a good time to taking a step back and looking at
> triple-level annotations with
>  fresh eyes once more?

POWDER?

>
>  - closing the XML-RDF gaps: It seems that the XML and RDF languages specing
> goes determined parallel
>   paths in W3C with no real visible perspective for convergence. Integrated
> XML+RDF query languages,
>   or maybe an XG for XML2RDF/RDF2XML would be cool: GRDDL is a starting
> point, but the current way to
>   translate into RDF/XML and then use XML transformation languages instead
> of directly using SPARQL/SPARQL-Update
>   seems like an unnecessary detour.

Think I might disagree there - so much of the XML stack is tied to a
tree model, and that distracts from the intuitive graph perspective on
RDF.

> In general, I think a workshop on "RDF what's next" or specific workshops to
> any of the above-mentioned topics would
> be necessary/welcome. The easy to solve issues (what anybody a lot of people
> are using, things like Turtle, named graphs)
> should probably go right into rec track, for the othe issues XGs might be a
> good instrument.

Absolutely - a lot of the annoyances we have could probably be bashed
out through a bit of informal f2f time, and from there getting the
rubber stamp would be a no-brainer.

> Nice, so now thatI've written that down I can wait for Chrismtmas and see
> what Santa Sandro will bring ;-)

Love it! Bet he's developing a white beard now...

Cheers,
Danny.

-- 
http://danny.ayers.name

Received on Thursday, 5 November 2009 00:27:57 UTC