Re: RDF namespace URIs [was: FW: RE: Jena in Haystack]

Ralph,

Please see my follow up message to www-rdf-dspace attempting to clarify this. My
original message was confused and confusing on this point I agree (written at
the end of a long Friday!) and all of the errors in here are mine - not the
working group's. 

Even though this was originally a private email exchange I should have been more
careful. When the exchange got posted to the public list I re-read it more
carefully, realized my error and posted clarification.

> I'll have to complain to the RDF Core WG then.   This
> heuristic is unreliable and unnecessary.  

Agreed. I was completely incorrect in ascribing this to the WG group here and
apologize. 

In jena, by default, you explicitly specify predicates using a qname pair - this
avoids any ambiguity. However, there is a buried piece of the API would does
allow you to take a URI and guess the namespace/localname split. You are, of
course, correct that this is simply best avoided by using explicit specification
of namespaces for example via isDefinedBy. 

> Let's encourage by example -- by writing code that uses
> explicit references to namespaces with a property -- and
> get the schemas to follow along.

Agreed.

> If jena believes that there is a URI that conforms to the URI RFC
> but that is disallowed by RDF then we have a serious problem.

No, jena does not believe that.

> I'll have to pay closer attention to the RDF Core WG discussion
> on this issue, I see.

No, I just have be more careful about what I write an email late on a Friday!
Apologies again.

Dave

Received on Friday, 19 July 2002 09:07:56 UTC