Expressing position in RDF

[Bernard Vatant]

Ed
> In a certain context I might actually say, "HDNNNN has RA 13.53" and 
> totally leave out epoch, equinox, system, units, and error and it 
> would be useful.  
Sure. And somehow going against a point often made by Semantic Web 
gurus, that assertions expressed as RDF triples are context-independent, 
since a URI is supposed to represent always and everywhere the same 
"thing" ...
> In a peer reviewed journal, I would not think of saying this unless  
> all  those  metadata are  clearly tied down.  So, most of the 
> discussion on this thread today is not a discussion of OWL/semantics 
> but rather of astronomical contexts and purposes.  
Agreed again. That was my point, actually, but I must admit it was too 
decorated with remarks some have called "philosophical". :-[
Well, too bad, I wanted to make a practical point : if something is 
complex, and you know the complexity, either you represent it, or if you 
simplify, try to make clear what, why and how you simplify. Having 
complexity handled by context is often a good idea for all sort of 
practical reasons. But all the point of the SemanticWeb is to get data 
usable out of context. So either data have to carry along their context 
in more or less convoluted and akward descriptions, or the context 
semantics has to be exposed somewhere and resolved if necessary ...
> We hit these same sticking points in STC and in DM.  The beauty of 
> OWL/semantics is the openness that allows it to grow until it becomes 
> completely parallel to normal language so you can provide whatever is 
> appropriate for the given context.  The downside, besides sheer size,  
> is some noninteroperability.   The VOEVENT software might prefer  
> "MDNNNN  a:hasRA  13.53^^float",  and that is fine. But, a  more 
> elaborate analysis package  might simply not accept such statements.  
> With some  intense coding, it may  understand that this statement sets 
> an RA in some equatorial system at some epoch, and it will either set 
> these as unknowns or put out a request for additional information.
> This may be a heretical statement, but complete interoperability may 
> not be the best course.
Why heretical? Semantic Web technologies do not aim at complete 
interoperability. One motto is : "interpret what you understand, ignore 
the rest". But instead of "ignore", you can indeed "ask for more". :-)

-- 

*Bernard Vatant
*Senior Consultant
Vocabulary & Data Engineering
Tel:       +33 (0) 971 488 459
Mail:     bernard.vatant at mondeca.com <mailto:bernard.vatant at mondeca.com>
----------------------------------------------------
*Mondeca**
*3, cité Nollez 75018 Paris France
Web:    www.mondeca.com <http://www.mondeca.com>
Blog:    Leçons de Choses <http://mondeca.wordpress.com/>
----------------------------------------------------**