W3C home > Mailing lists > Public > semantic-web@w3.org > September 2008

Re: SWIG F2F during W3C TPAC week, Oct 20/21 (Cannes, France)

From: Peter F. Patel-Schneider <pfps@research.bell-labs.com>
Date: Wed, 03 Sep 2008 13:13:45 -0400 (EDT)
Message-Id: <20080903.131345.71447437.pfps@research.bell-labs.com>
To: danny.ayers@gmail.com
Cc: michael.hausenblas@joanneum.at, danbri@danbri.org, semantic-web@w3.org, ivan@w3.org, parcher@fosi.org

From: "Danny Ayers" <danny.ayers@gmail.com>
Subject: Re: SWIG F2F during W3C TPAC week, Oct 20/21 (Cannes, France)
Date: Wed, 3 Sep 2008 18:31:54 +0200

> 2008/9/1 Peter F. Patel-Schneider <pfps@research.bell-labs.com>:
> >
> > From: "Hausenblas, Michael" <michael.hausenblas@joanneum.at>
> >> Peter,
> >>
> >> >I would be very interested in reading a full specification of
> >> >N3.  Could you point me to one?
> best I know of is:
> http://www.w3.org/2000/10/swap/cwm.tar.gz
> (although I'm not sure about literal handling)

Hmm.  That appears to be the source for CWM.  Which file should I go to
in the tarball?  Is there any formal relationship between CWM and N3?

Following links from files in the tarball ends up at 
which does claim to be "the specification of the Notation3 language".
The document appears to be the source of a large portion of the team
submission on N3.

> >> Dunno, is http://www.w3.org/TeamSubmission/n3/ 'full' enough?
> > In a word, "non".  In two words, ....  (You probably don't want to know.)
> >
> > Can I take that document and reliably implement N3?
> Try it. Who knows? ;-)

My view is that the document is horribly inadequate as a specification
of N3 as a representation language.

> > Perhaps I could
> > implement a tool that takes an N3 document and produces a related RDF
> > graph although I'm doubtful of even that.
> I would imagine that depends which constructs of N3 are used. (I
> believe in general you'd get multiple graphs + rules)

Oh?  This is a bit surprising.  I had thought that an N3 document could
be processed into a single RDF graph.  I would be interested to see
where the multiplicity comes from.

> > However, I believe that there
> > is no chance that I (or anyone else) could use solely that document to
> > implement reliable reasoning in N3 or to develop a formal meaning for
> > N3.
> I've not spent much time with N3 proper, but it does strike me as
> worthwhile exploration. Also I hate to have to contradict you, but I'm
> sure you personally could develop a formal meaning for N3 (i.e. fill
> in the blanks), and I've no doubt the creators of N3 would be grateful
> for it to appear in the next rev of the spec.

Nope, I actually wouldn't know where to end, or even, really, to begin.  

Of course, I *could* implement something that might look a bit like N3,
if I made a whole bunch of assumptions about the logical underpinning of
N3.  But that is not the way things should work.  The N3 document should
provide these underpinnings to me.

> But Michael, Phil, Peter, whoever started this thread - is there any
> particular reason you need N3 over the Turtle subset (plus maybe
> SPARQL's notion of named graphs)?
> http://www.w3.org/TeamSubmission/turtle/

Who knows?  Not I.  I didn't start the thread, I was just responding to
the claim by Phil Archer in 
that N3 is "fully specified and understood".

> Cheers,
> Danny.

Received on Wednesday, 3 September 2008 17:15:20 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:49:44 UTC