3 Feb
2004
3 Feb
'04
6 a.m.
On Mon, Feb 02, 2004 at 01:46:31 -0500, Paul Winkler wrote:
On Mon, Feb 02, 2004 at 06:42:47PM +0000, Steve Harris
wrote:
How is this done in the RDF world?
You create a mapping schema which says
GreenThings rdfs:subClassOf ColouredThings .
RedThings rdfs:subClassOf ColouredThings .
(and the same for the properties that need translating.
So any queries you have on GreenThings will be satisfied by instances of
ColouredThings - not a perfect match as you can see, but better than
nothing.
A better example is if someone gives you a load of data in Dublin Core you
can say
dc:title rdfs:subPropertyOf my:name .
dc:identifier rdfs:subPropertyOf my:url .
...
So you will get eqivalents in your schema for all the DC properties. If
you reverse it as well you will get mapping in both directions so people
who use DC can understand your data.
Splitting
and joining classes? You lost me there.
Refactoring?
Yeah, I guess. Eg. if you had Green Things and Red Things and you wanted
to replace them with Coloured Things. sorry to be a pest, it's just new concepts to me
and I'm slow
to grasp the benefits.
Not at all, its a bit of an odd concept to get your head round. Someone
desciribed ontologies (things like RDFS) as an equivalent of an API for
data - which is close enough.
- Steve