W3C home > Mailing lists > Public > semantic-web@w3.org > November 2006

Re: geo:lat owl:samePropertyAs vcard:latitude? (vCard/RDF)

From: Norman Gray <norman@astro.gla.ac.uk>
Date: Fri, 24 Nov 2006 12:50:57 +0000
Message-Id: <EF1119C8-0286-400F-9E12-68AE7252E2F5@astro.gla.ac.uk>
Cc: semantic-web@w3.org
To: Bernard Vatant <bernard.vatant@mondeca.com>


Hello Bernard,

On 2006 Nov 24 , at 08.44, Bernard Vatant wrote:

> Actually this use case should help us to set up best practices for  
> any similar situation where you have to deal with numerical  
> properties likely to come out of different measurement systems,  
> hence with different values at a level of precision needed for  
> technical applications, such differences being irrelevant in less  
> technical contexts.

What you say is true, and there are long, _long_, arguments one can  
have about how to represent data with this sort of precision.   
However, I'm not convinced those arguments are always worth having.

The arguments I'm thinking of are in the context of the Virtual  
Observatory, and how you express astronomical data in an  
interoperable and reusable way.  The applications in question are  
more elaborate than mashups, but there's still an acknowledgement  
that with interoperability comes some manageable degree of  
imprecision.  If you're aware of that, it needn't stop you doing good  
science.

More precision is generally better, but there comes a point where the  
interoperability standard becomes more complicated than the standards  
it's wrapping.

> For positions of stars, whatever catalogue is good for the backyard  
> astronomer, whereas the professional use would certainly want to  
> distinguish Hipparcos data from Gaia's when they are available.

Gaia vs. Hipparcos is a good example, but there will be analogues in  
other areas.  If you know enough to care about the difference, then  
you know enough to be able to use the underlying data directly, and  
you're probably making sufficient investments of time that the  
wrapping is saving you nothing.  We've moved rather beyond lat/long  
here....

I suppose the short version is just that consensus is extremely  
expensive, and not always cost-effective.

All the best,

Norman


-- 
------------------------------------------------------------------------ 
----
Norman Gray  /  http://nxg.me.uk
eurovotech.org  /  University of Leicester, UK
Received on Friday, 24 November 2006 13:10:33 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 21:45:12 GMT