W3C home > Mailing lists > Public > w3c-rdfcore-wg@w3.org > October 2001

Re: big issue (2001-09-28#13)

From: Jan Grant <Jan.Grant@bristol.ac.uk>
Date: Mon, 1 Oct 2001 17:04:18 +0100 (BST)
To: Jeremy Carroll <jjc@hplb.hpl.hp.com>
cc: w3c-rdfcore-wg <w3c-rdfcore-wg@w3.org>
Message-ID: <Pine.GSO.4.31.0110011653560.8265-100000@mail.ilrt.bris.ac.uk>
On Mon, 1 Oct 2001, Jeremy Carroll wrote:

> I am less than comfortable with the direction of
> Graham's
> http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2001Oct/0002.html
> Jan's
> http://ioctl.org/rdf/literals
> which both seem to make much more of a Literal than M&S.
> Our job is constrained by the charter.
> I see our job as making the best sense we can out of M&S and schema, and not
> reinventing RDF from the ground up.
> I include in our job that of finishing various half finished items in M&S
> and RDF schema, but not going completely against the spirit of the two
> documents.

M+S para 203, Schema para 123 both indicate that there are blanks to
fill in in the current definition of literals.

> Given that these documents are problematic "making the best sense we can"
> leaves us a lot of room. I also note that data typing, and specifically the
> relationship between RDF and XML schema datatypes, probably should be
> addressed in part in this round.
> I hope we can do that in a way that:
> + is sufficient for what the community needs now,
> + does not break too many older tools that have made a bona fide attempt to
> implement M&S and schema
> + leaves the future sufficiently open for an RDF 2 WG.
> We do not have a blank piece of paper.

I agree, which is why I almost put "... and I suspect this all gets left
until RDF 2" in the literals doc you pointed to.

However, I'm a little surprised since what I suggested can be done with
as little as (say) _representing_ a literal as (unicode + uri) pair; in
particular, we don't need to require RDF parser writers to support an
open-ended set of data types.

It comes down to this: if you want datatyping on literals, you have two
choices (or rather, two classes of choices).

1. Have typed literals (the more radical proposal, but can still be made
reasonably compatible with existing parsers)

2. Have literal = string (+ associated language tagging etc.)

In this case you're forced to do datatyping with the
	<a> <foo> _:a .
	_:a <rdf:type> <blah> .
	_:a <rdf:value> "bletch" .
construct, or an analagous one.

Personally I find the latter construct somewhat ugly; that's not to say
that it's wrong, it's just making do with what RDF gives you. There also
seems to be a bit of effort going into producing RDF "databases";
personally, I like my databases to know about data types - however, this
might not be "the RDF way" :-)


jan grant, ILRT, University of Bristol. http://www.ilrt.bris.ac.uk/
Tel +44(0)117 9287088 Fax +44 (0)117 9287112 RFC822 jan.grant@bris.ac.uk
"My army boots contain everything not in them." - Russell's pair o' Docs.
Received on Monday, 1 October 2001 12:04:47 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 20:24:05 UTC