W3C home > Mailing lists > Public > www-rdf-interest@w3.org > November 2004

Re: I guess it's a stupid questions...

From: Daniel O'Connor <daniel.oconnor@gmail.com>
Date: Sat, 13 Nov 2004 20:39:47 +1030
Message-ID: <106cc1200411130209c37d53a@mail.gmail.com>
To: www-rdf-interest@w3.org

>Be careful here. "Meaning" here is nothing more than a set of rules,
connecting different names like "person", "box", "bob".

Don't be pedantic :)
You're scaring away the fish. It was just a simple way to demonstrate
some of the inbuilt features of RDF as opposed to those of XML - the
less complexity the easier it is to understand a concept.

>First, you need units - so each box side is a 'double' - not a triple...

Actually,
"box width 1" is a triple.
"box height 1" is another triple.

>Second - chosing a box is a bit confusing since it has (conveniently) three
dimensions. Does that mean we need a double to decribe a rectangle - and a
n-iple (no offence intended) for a hypercube?

Nope. Think of how you can express complex things in programming -
while loops, polymorphism, objects, etc. When push comes to shove,
that's all a long string of 0's and 1's.
A triple is a very decent base which you can use to build more complex
things out of.

>Also, lets say we are interested in the specific gravity of the water in
bob's box, at a given temperature whose dimensions for some strange reason
have been variously recorded in inches, kilometers and cubits.

That's an issue yes. But (in OWL), you can change about a few
namespaces in the xml document, make an UnderWaterBox a subclass of
Box, and still be ok.
Or, you could go really distributed, and make all of the values to be URI's.

AFAIK there's no easy way to say "this class equates to adding
together 5 other classes" or other math type operations.
I ran into the problem myself when trying to *easily* author a cooking
ontology - saying a teaspoon was 15ml's in volume, a litre was 1000,
etc.

>Alternatively, by leveraging the widely deployed FishBoxML 'standard', a
more straightforward analytical approach can be used - maybe... ;-)
Heh, perhaps. But then you get the mess of RSS. And Atom. 

Take a look at http://www.foaf-project.org/ perhaps. There's a way to
talk about people and who they know, and it's not broken yet - because
adding in or forgetting elements from a FOAF description of a person
doesn't really matter.
When you want to talk about something in more detail you add in
another namespace and off you go.

RSS, however (excluding 1.0), is a mess. If I wanted to use RSS to
describe what I am listening to (see http://www.audioscrobbler.com/),
I have to accept that I'm producing a list of Links with Titles and
Descriptions - instead of Musical Songs.

http://www.ahsonline.com.au/audioscrobbler/as_recent.rdf is some
sample RDF I'm doing for audioscrobbler.com - It's basically valid RDF
(RSS 1.0) which contains a lot more information than an RSS 2.0 feed
could ever. It describes a Person, Tracks, and Artists.

(Note, it's a work in progress, so probably slightly broken hehe)

-Daniel O'Connor
-- 
http://www.ahsonline.com.au/dod/FOAF.rdf
Received on Saturday, 13 November 2004 10:09:48 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 5 February 2014 07:14:56 UTC