W3C home > Mailing lists > Public > www-archive@w3.org > December 2001

Re: review of XML in 10 points [was: AGENDA...]

From: Dan Brickley <danbri@w3.org>
Date: Fri, 7 Dec 2001 09:58:26 -0500 (EST)
To: "Peter F. Patel-Schneider" <pfps@research.bell-labs.com>
cc: <connolly@w3.org>, <phayes@ai.uwf.edu>, <janet@w3.org>, <bert@w3.org>, <em@w3.org>, <liam@w3.org>, <www-archive@w3.org>
Message-ID: <Pine.LNX.4.30.0112070917220.14696-100000@tux.w3.org>

On Fri, 7 Dec 2001, Peter F. Patel-Schneider wrote:

> Sure.  Except .... I don't have time, I don't have support from my
> management for doing it, and I hate writing primer-like materials.
> 	:-)

(Maybe we can pinch bits of this text for the Primer anyway...? eric?)

> > Anyway, nitpicking (with an eye to the rdf primer as target doc):
> >
> > 	'true knowledge' - is there any other kind (apart from
> > 	non-propositional, which rdf/webont doesn't really
> > 	attempt to capture...)
> I was using true as in real, as opposed to fake.  I admit that I also used
> it partly because of how it made the phrase sound, and also partly because
> of the other meaning of true (i.e., not false).

heh :)

I expect RDF/WebOnt to be used extensively for lies and half-truths, as
per most email/internet/web precedents. When I get my first RDF/WOL spam
message, I'll know we've gotten somewhere! (and yes -- picking up on
another thread --  RDF reification isn't the best way to represent and
reason about that kind of content...).

> > 	'collections of meanings' - I prefer to avoid physicalist
> > 	metaphors for meaning. Talking about meaning as a type of thing or
> > 	a type of stuff only confuses. You can't count it; and you can't
> > 	weigh it either... But you can count (XML-encoded) claims about
> > 	the proper/appropriate use of vocabulary. Meaning itself isn't
> > 	downloadable, collectable... only (as per the final sentence)
> > 	representable.
> As a formalist (or at least a quasi-formalist) I try to use neutral terms
> (like ``collection'') when I don't want to invoke any of the aspects of
> more-specialized terms.  It might be better here to use a bare plural
> (``meanings'') or even a bare singular (``meaning'').

As a scruffy ex-philosopher type, I suppose my tendency is to avoid nouns
where their use might spawn "angels on pinheads" discussion (eg. "what is
a 'Meaning'?" etc.). So I'd tweak your text slightly, eg as follows:

> > >	 Collections of meanings for
> > > terms in a certain area (from "shopping" to "manufacturing") are called
> > > ontologies and are a necessary part of the Semantic Web.

  Descriptions of the meaning of terms in a certain area (from "shopping" to
  "manufacturing") are called ontologies and are an important component of
  the Semantic Web.

(this does slight violence to the way folk in this field use 'ontology',
but I can live with that)

I'm also inclined to seek words that gloss 'meaning' in terms of
constraints on (sensible) usage. For eg:

  	While it is difficult (*undestatement*!) to represent the full
	meaning of such terms in a machine-processable way, we
	can help machines make use of Web ontologies through giving them
	simple rules for the proper use and combination of these terms. To
	understand natural language terms is in large part to understand
	the constraints on their sensible usage; much the same is also
	true of Web ontologies. [eg here]

Having written that, I'm not entirely convinced by it. Needs an example at
least. Using words like 'constraint' can cause confusion eg with XML
notions of validation; the RDFS spec currently suffers from that (though
next draft won't).

> > Grumbles aside, this is useful text...
> Thanks.  I'm not sure what the next step should be.

Me neither. Eric? where are things up to with the Primer?


Received on Friday, 7 December 2001 09:58:36 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 22:31:39 UTC