Re: RDF Triples Transport

(tab separated, ugh: Why not just use ASCII instead of Unicode while
you're at it?)

The <Statement> technique is tempting -- I started a writeup a while back
with the idea of it being a potential note. I was stumped when it came to
dealing with anonymous nodes; there are several conventions we might adopt
and none seemed particularly appealing or obvious. Also we'd need to do
namespace declarations and language tagging properly...

Dan

ps. if anyone has photos etc from the IG meeting last week at a public
url, let me know -- I'm working on the writeup

On Tue, 6 Mar 2001, Seth Russell wrote:

> Why not just use a tab delimited file?
>
> subject TAB predicate TAB object CRLF
>
> It's far simpler and doesn't require anybody to write a parser, and the
> generator is triveal too.
>
> Seth
>
> ----- Original Message -----
> From: "Aaron Swartz" <aswartz@upclink.com>
> To: "RDF Interest" <www-rdf-interest@w3.org>
> Cc: <barstow@w3.org>; "Dave Beckett" <dave.beckett@bristol.ac.uk>
> Sent: Tuesday, March 06, 2001 7:26 PM
> Subject: RDF Triples Transport
>
>
> > At the RDF IG F2F there was some discussion about ways to move pure
> triples
> > around from parser to parser. Some of us decided the best way to do it
> would
> > be to have a set of reified statements, like so:
> >
> > <rdf:RDF ...>
> > <rdf:Statement>
> >   <rdf:subject rdf:resource="..." />
> >   <rdf:predicate rdf:resource="..." />
> >   <rdf:object rdf:resource="..." />
> > </rdf:Statement>
> > ...
> > </rdf:RDF>
> >
> > This format can be both parsed as RDF and as plain XML, so it's very
> simple.
> >
> > As a demonstration of this, I have put up an web service that will take a
> > URI, run it through SiRPAC and stick out a bunch of triples:
> >
> > http://purl.org/swag/rdf2triples?url=(url-goes-here)
> >
> > I'd love to see others put up similar interfaces to their parsers. (Eric
> > dares you to. ;-)) The idea is that then we can begin to build
> cross-parser
> > apps on top of this. For example, the SiRPAC servlet could offer a choice
> > between various parsers, and Dave Beckett's parser test suite could run
> > itself.
> >
> > Another application in the back of my mind is using this for signing sets
> of
> > triples. The major problem with this however is some way of ordering the
> > statements.
> >
> > Anyway, some food for thought.
> > --
> > Aaron Swartz <me@aaronsw.com>|  SWAG: Semantic Web Agreement Group
> >   <http://www.aaronsw.com>   |       <http://purl.org/swag/>
> > AIM: JediOfPi | ICQ: 33158237|     helping build the next web
> >
> >
>

Received on Tuesday, 6 March 2001 22:44:12 UTC