W3C home > Mailing lists > Public > w3c-rdfcore-wg@w3.org > December 2002

Re: specifying literals

From: Dan Connolly <connolly@w3.org>
Date: 15 Dec 2002 20:39:32 -0600
To: Graham Klyne <GK@NineByNine.org>
Cc: Jeremy Carroll <jjc@hplb.hpl.hp.com>, w3c-rdfcore-wg@w3.org
Message-Id: <1040006371.7171.9.camel@dirk.dm93.org>

On Sun, 2002-12-15 at 04:18, Graham Klyne wrote:
> At 06:13 PM 12/13/02 -0600, Dan Connolly wrote:
> 
> >I see yet another specification of literals in
> >http://lists.w3.org/Archives/Public/www-archive/2002Dec/att-0053/00-rc#section-Graph-Literal
> >
> >That has the same problem.
> >
> >Here's a suggested replacement:
> >
> >6.5 RDF Literals
> >
> >   To define a literal in an RDF graph,
> >   let S be the set of Unicode
> >   strings in Normal Form C,
> >   L be the set of language identifiers [RFC3066],
> >   and U be the set of RDF URI references.
> >
> >   A plain literal is an element of the
> >   union of S with SxL; i.e. it's either
> >   a string or a string paired with a
> >   language identifier.
> >
> >   A typed literal is an element of the
> >   union of SxU with SxUxL.
> >
> >   Note that U and L are disjoint
> >   (every member of U contains a colon;
> >    no member of L contains a colon),
> >   so SxL doesn't intersect SxU.
> 
> Eh?  Where's this about literal strings not containing colons come from?

I think you've confused S with L.

"let ... L be the set of language identifiers [RFC3066]"

> As far as I'm concerned, "a:b" is a perfectly good literal.

but it's no good as a language identifier, is it?
[I have to admit I didn't check closely.]

> >Then literal equality falls out from
> >the traditional definition of tuple
> >equality and string equality.
> >
> >I don't see what the NOTE about literals
> >being distinguisable from URI references
> >is supposed to mean; that's the same sort
> >of double-speak that's giving us trouble
> >in the XML Schema spec.
> 
> It's meant to mean that:
> 
>    ex:subj ex:prop <http://example.org/abc> .
> 
> and
> 
>    ex:subj ex:prop "http://example.org/abc" .
> 
> are different graphs.

Those are different graphs because
"http://example.org/abc" and
<http://example.org/abc> are different terms,
not because the URI http://example.org/abc
is different from the string 'http://example.org/abc'.
Hmm... I guess that means we can't let
the URI and the string-literal be terms
themselves... they do need to be sorta
wrapped in something, syntactically...

Maybe I'm barking up the wrong
tree here... I need what the term
"http://example.org/abc" to *denote*
a unicode string, but I don't need
the term itself to *be* a unicode string.

But somewhere there's a semantic consition
that literal terms denote themselves, no?

I guess I'll have to think about it some more...

-- 
Dan Connolly, W3C http://www.w3.org/People/Connolly/
Received on Sunday, 15 December 2002 21:40:50 EST

This archive was generated by hypermail pre-2.1.9 : Wednesday, 3 September 2003 09:54:55 EDT