RE: [Fwd: xmlns, uri+name pairs or just uris..? Clarification n eeded.]

Pierre Antoine,

Thanks for the clarification; it is clear what you mean now.

I have been assuming that an RDF processor that wanted to fetch
schemas dynamically when it needed them would need some function
which could take the URI and return the schema.  This function
may of course fail and as you say the software must go on as 
best it can.  I had assumed that this function could work on URN's
as well as URI's.

> What I wanted to say is:
> although reasonable, that asumption is too strong. URIs 
> (identifers) differ from URLs (locators) in that: they do not 
> *always* allow to retrieve the correpsonding resource. If 
> they don't, the application should work anyway. If they do, 
> very well, it will work better !
> 
> By restricting the property URIs to contain the URI of their 
> defining schemas, RDF restricts itself to "friendly" URIs, 
> which is IMHO a loss of generality.
> Hope this is clearer.
> 
> > If an application can
> > clearly distinguish between the original model and the annotions
> > then, aside from implementation convenience, there should be no
> > problem.
> 
> I believe that the property rdf:isDefinedBy is, by 
> definition, a meta-property.
> So there should be no problem with Dan's proposition of 
> generating additional statements with that property.

One of the properties I'm expecting an implementation to have is
the ability to round trip, i.e. to be able to serialize a model 
to xml, read it back in through a parser and end up with exactly
the original model.  I don't think I'm going to be able to
accomplish that if the parser is adding statements that were
not in the original model.

And now, is Graham going to say - "That's not a reasonable 
requirement."?  :)  Maybe its not, and its a style thing.  One
would have to be careful with digital signature algorithms though
if a parser can modify a model on the way through.

Brian McBride
HPLabs

Received on Tuesday, 1 August 2000 13:26:00 UTC