SV: do we need it?

[Frederic V. Hessman]

> Are applications likely to use the SV directly or via, say, data  
> models. I can see a data model item having a pointer to the term/ 
> concept that it refers to in the SV and apps using that to present  
> a definition to the user, but would any apps be likely to access  
> the SV directly?
I sure hope so.   Most astronomers aren't directly interested in  
theoretical ontology studies but labeling, searching, identifying in  
human understandable terms.    No, they shouldn't see  
"star.variable.cataclysmicVariable" if they are more happy with "CV",  
but the app will have to decide what human layer is appropriate and  
pass the invisible token off to it's computer colleagues.

> > underscores (eg gamma_ray) will turn out to be a terrible choice and
> > maybe camelback (eg GammaRay) is even worse (probably it doesn't
>
> If the terms in the SV are URIs, say http://ivoa.net/SV/ 
> cosmicBackground (to use the format from Andrea's document), then  
> they can remain the same while the terms they reference, ie  
> everything a user will be shown, can change. In which case the  
> decision over format is ours.
>
> I would argue against the SV containing any structure though, so  
> use cosmicBackground and not cosmology.cosmicBackground.
As stated in the draft working draft, the suggested ontological  
structure was purposefully and only used to make the creation of the  
list of tokens more manageble for humans.  Got a new token?  It's  
easier to go to the "cosmology" group (whatever the group represents)  
and look to see if it's already there than to go looking for all  
possible equivalents:

...
cosmeticDefects
cosmicBackground
costEstimate
...

If the tokens are bare-bones, then we have to be careful to have good  
thesaurus tools which insure that we don't accidentally double or  
triple the entries, thereby making the vocabulary useless.

> No, broader and narrower are relationships: so you would record the  
> relationship that SeyfertGalaxy is a narrowing of the term, Galaxy.  
> Certainly we would want to record both terms, but do we want to  
> record the realtionships? And, to start with, I'd say no. The above  
> relationship is probably a safe bet but many (and I'm thinking of  
> perhaps cosmology -> cosmicBackground) are less certain or more  
> problematic, so would reckon that we skip all relationships at the  
> beginning.
>
> We might even, later, decide that we'd rather use OWL or similar to  
> record relationships rather than SKOS.
If the fundamental purpose of the vocabulary is to provide tokens,  
then one doesn't need any further info than the token, the labels,  
and the definition.   However, it will be very nice for apps to be  
given a minimum set of un-official ontological hints that formally  
don't constitute a real ontology but sure would be handy.   Assume  
you're writing an query interface and your clients want to ask about  
observations of various galaxies: the SV et al. could immediately  
give you a handy, official, standardized list of types of galaxies  
all prepared and ready to go - but only if the computer can figure  
out which tokens might have anything to do with galaxies.   Do we  
really want to wait for the ultimate ontology in order to do this  
trivial task?

Rick

------------------------------------------------------------------------ 
------------------------
Dr. Frederic V. Hessman     Hessman at Astro.physik.Uni-Goettingen.DE
Institut für Astrophysik          Tel.  +49-551-39-5052
Friedrich-Hund-Platz 1         Fax +49-551-39-5043
37077 Goettingen                 Room F04-133
http://www.Astro.physik.Uni-Goettingen.de/~hessman
------------------------------------------------------------------------ 
-------------------------
MONET: a MOnitoring NEtwork of Telescopes
http://monet.Uni-Goettingen.de
------------------------------------------------------------------------ 
-------------------------


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.ivoa.net/pipermail/semantics/attachments/20070919/8245cdb5/attachment-0001.html>