- From: Dan Brickley <danbri@w3.org>
- Date: Mon, 21 Apr 2003 20:34:12 -0400
- To: "Roger L. Costello" <costello@mitre.org>
- Cc: www-rdf-interest@w3.org
One quick review comment:
(jumping straight to the nitpicking!)
re http://www.xfront.com/owl-quick-intro/sld002.htm
Slide 2 is a little inaccurate in it's first bullet, in that you ascribe
to OWL characteristics that belong to the broader RDF approach. Specifically,
RDF is concerned with the instance level and with logical relations, OWL
provides supporting machinery for making richer generalisations, eg.
criteria for class membership, relationships amongst classes etc. It's a
bit confusing to describe OWL solely as 'an XML vocabulary' since it doesn't
really work in standalone fashion; it's part of the broader framework we
began with the RDF syntax, graph data model and schema language, and which
we may continue with work on rules, query etc etc.
Perhaps:
(thrown to the mailing list for discussion, shredding etc :)
* OWL is an RDF/XML vocabulary for describing the kinds of categories
and logical relationships used in RDF/XML data exchange on the Web.
* OWL extends the Resource Description Framework by providing some
formally-defined mechanisms for stating generalisations on the Web.
- A collection of definitions for a domain is called an "Ontology"
...
This costs you an extra bullet point and acronym, but perhaps things could be
compacted. I think introducing OWL as if it sat directly atop XML without
RDF being acknowledged as a named architectural component is an approach
worth considering, as is the approach of calling all of this stuff 'RDF'.
Currently the specs use 'RDF' for the basic graph approach and syntax, and
'OWL' for the Ontology layer. It's probably best if OWL and RDF intros
followed the specs in this, although I welcome discussion on
www-rdf-interest as to whether this acronym soup could be better
presented to users...
Another technical question re slide 20:
http://www.xfront.com/owl-quick-intro/sld020.htm re "a camera with a
75-300mm zoom lens", is that supposed to indicate a discrete value, or do
zoom lenses take values such as 75mm, 200mm etc?
Something of an aside that I might announce properly when more polished,
but which relates somewhat to your device-description use case, I spent
some time today working with some UAProf / CC/PP device descriptions in RDF,
working notes on which can be found at
http://esw.w3.org/topic/UAProfIndex in the ESW wiki. I harvested device
descriptions from http://w3development.de/rdf/uaprof_repository/ and loaded
them into an RDF store, trying queries such as
[[
SELECT ?model, ?vendor, ?size, ?doescolor, ?doesimages, ?doestextin, ?hwp,
WHERE
(prf::Model ?hwp ?model)
(prf::Vendor ?hwp ?vendor)
(prf::ScreenSize ?hwp ?size)
(prf::ColorCapable ?hwp ?doescolor)
(prf::ImageCapable ?hwp ?doesimages)
(prf::TextInputCapable ?hwp ?doestextin)
USING
prf for http://www.wapforum.org/profiles/UAPROF/ccppschema-20010430#
]]
...this is more for mobile phones than cameras, but I suspect
http://www.w3.org/TR/di-princ/ and nearby might be of interest. After I
built an aggregation of these descriptions the first thought I had was
that with a bit more additional metadata they'd be quite good for
building an online store, eg. 'find me something with such'n'so
capabilities, color, screensize etc of type s:CameraPhone that works in
the UK and costs less that 200ukp...
Dan
Received on Monday, 21 April 2003 20:34:13 UTC