Re: rdf inclusion

On Tue, 23 Apr 2002, Jeff Heflin wrote:

> Tim,
>
> If you refer to a term in an ontology without importing then I suggest
> that you are using a name without agreeing to its definition. This is
> not an error condition, but it may mean that DAML+OIL-compliant agents
> won't conclude everything you intend them to.

This is going to happen a *lot*. Many terms (for some sense of that word)
in xml/rdf won't be classes and properties, just plain old URI references.

Is using a class/property name without 'agreeing to its definition' any
worse than using some abitrary URI without similar agreement? (Can't we
all just muddle along...?).

Where does this stop? For example, consider a class
http://example.com/xmlns/Bachelor defined as an unmarried man. Presumably
to agree to the definition of 'Bachelor', we'd need to use the terms
'unmarried' and 'man'. And using those without agreeing to their
definition would be... bad? So we'd need to agree to the definitions of
all the terms used in their defintions... and so on.

Dan


-- 
mailto:danbri@w3.org
http://www.w3.org/People/DanBri/

Received on Wednesday, 24 April 2002 19:36:32 UTC