Vocabularies and VOEvent
Frederic V. Hessman
Hessman at Astro.physik.Uni-Goettingen.de
Tue Feb 5 06:19:45 PST 2008
Can't help but add my 2 bits:
On 4 Feb 2008, at 11:10 pm, Roy Williams wrote:
> The vocabulary effort is very exciting, there are a lot of messages.
> But I am still waiting for the cost benefit analysis. The number of
> emails sent to this group shows us the cost. But, for myself at
> least, I am torn between the two approaches to defining "Why" this
> event may have occurred, asking if the benefits of the higher cost
> version outweighs its costs. The older I get, the more I want to
> keep with KISS (Keep It Simple and Sufficient),
> and the less I am interested in finding the general solution.
> Anyway, here are the choices for VOEvent 2.0 as I see them:
The only sign of hope, in this regard, is the comforting knowledge
that there IS no general solution.
> (1) The KISS approach:
> We could follow the astronomer's telegram (*) and define a short
> list of words to be used to describe event hypothesis, such as those
> below, with "check all those that apply". The query and API is simple:
> select * where event.why == "Novae" .....
> if eventType == "Novae" then .....
> If this is not sufficiently precise, I would like to know the
> plausible, real-life situation, where such imprecision will become a
> burden. In
Done problem. Ages ago, I sent around a list of handy-dandy IAU/
IVOAT words which covered all thinkable terms like this and a bit
more. Pick and choose, knowing that the rest of us will be able to
know what you are saying. Norman's offer to add the AT vocabulary to
our list of examples is a great idea (one more ready-to-use-example to
> (2) The vocabulary approach:
> I know very little about it except that precision of expression is
> possible (eg "I want Supernovae Ia but not Ic"). I would like to be
> disembarrassed of my horrible prejudices. I worry that there are
> many vocabularies with no agreement. I worry that there are many
> representation technologies for vocabularies that use complicated
> languages. I worry that use of vocabularies will require knowledge
> of complex matters like namespaces, schema extension, sparql, skos,
> rdf, owl, etc etc. I worry about no being able to use SQL or even
> Xquery but must learn a new query language. I worry that the creator
> of the VOEvent will also need to understand these complex
> technologies, meaning that few event streams will be done this way.
> Are any of these valid worries?
Same solution, just that Roy thought his KISS vocabulary was
intrinsically different from a vocabulary we've expressed in SKOS,
whereas this is just a trivial question of simple format, a naming
convention, and the question of publication. Indeed, the issues
addressed by the original VOEvent protocol were MUCH more complicated.
This point obviously hasn't sunk in so I'll repeat it for the email
we're simply talking about puting the words in our vocabularies in a
standardized format, giving them names, and publishing them over the
internet. That's it - there's no magic beyond that: create the list
name it, and publish it so that we can all use it. This isn't rocket-
this isn't even bicycle-science, this is <paperClip/>-science.
> I do not want to pooh-pooh this promising vocabulary effort. It's
> just that I have not seen the reasons why (2) is so much better
> than (1) that the effort is justified.
The beauty of the IVOA/SKOS proposal is that Roy can have his semantic
cake and eat it too. If VOEvent wants it's own vocabulary, then by
all means - define one that you're happy with and don't wait for the
IVOA/IAU or anyone else to force feed what they think you need, but at
least write it down and publish it so that everyone else knows what
you're talking about.
More information about the semantics