What use the AstroOntology

[Kjetil Kjernsmo]

On Sunday 04 March 2007 12:19, Tony Linde wrote:
> We have to ask what we want to use an ontology for. Then decide if
> the structures of OWL are up to the job. The QSO/AGN discussion seems
> to show that any definitions added to the ontology need to be limited
> and for guidance only: they cannot serve as identification rules.
>
> At least, I don't think they can. Can they?

Right, I started to think along the same lines. I don't quite know in 
the general case, for AGNs, where I have my (somewhat dated) expertise, 
I think it would be difficult. 

Also, I think that given the distributed nature of the Semantic Web, 
each community can trivially add triples to the ontology as needed, so 
if AGN community finds that it can, it can very easily do so. 

Thus, when designing an ontology that spans all of astronomy, I think it 
sounds reasonable to create classes for object types that are in use, 
and as Anita noted, classes for objects types that have been in use 
relatively recently will also be useful in practice. 

In many cases, the distinction between different classes is blurry and 
may be subject to controversy and active research, in which case I 
think it sounds reasonable to only include class axioms (e.g. 
subclassing, disjointwith) where there is overwhelming consensus.

If a research community finds that more detailed classification is 
achievable, they can trivially do so by adding class axioms themselves. 
This way, making detailed statements will be left to those with 
expertise in a given field. If they find that they are able to 
formulate identification rules, great.

Moreover, in the case of controversy, if some researchers finds that 
adding controversial statements servers their purpose as it helps them 
use OWL tools, then fine, they can do that too, nobody else will be 
compelled to add those statements if they don't want to.

Best,

Kjetil
-- 
Kjetil Kjernsmo
Programmer / Astrophysicist / Ski-orienteer / Orienteer / Mountaineer
kjetil at kjernsmo.net
Homepage: http://www.kjetil.kjernsmo.net/     OpenPGP KeyID: 6A6A0BBC