Hi,
Since I will not be personally present when many of you meet in Europe
(now), I decided to send in my two cents to the group.
Over the last few months, talking to several VO people from US and Europe, it is clear that everyone agrees that something like the UCDs is great. Not everyone agrees on (1) the number of UCDs, (2) the hierarchical nature they are currently arranged in, (3) mechanisms of adding UCDs, ...
Suggestions range from leave UCDs as they are to completely replace them with alternate mechanisms (some of which have at least partially been thought of).
Having worked with UCDs for making Topic Maps based on UCDs in catalogs
(examples at http://www.astro.caltech.edu/~aam/science/topicmaps/ucd.html
- a specialised application based on these will be ready soon),
I have come to like the ease with which one can use UCDs to build
structure around them (though they certainly have their limitations, as
is seen by some "gaps" in some of the topic maps). I have not been
required to deal with the entire set of UCDs (e.g. Atomic Data, Orbit
etc. related UCDs) as I have not dealt with all kinds of catalogs. But
the hierarchical structure seems very useable. Having error terms on
various individual UCDs will be good and some work could go in there. Of
the 1400 or so UCDs, the 100 most accessed catalogs at Vizier use only
about 250 UCDs in all.
Allowing people to make their own UCDs (just like standards should not be a plural, not everyone should create their own Universals) will bring us back to square one even if what the personal UCDs mean is well documented. My R magnitude UCD is GunnR and some one elses is GR (both having used the Thuan-Gunn system). Okay so far. But then I also use GR to mean Gamma Ratio (whatever that is), and all hell breaks loose. A lot of time may have to be spent in writing more intelligent software to deal with these issues which could in the first place have been avoided.
So should there be "UCD police" who decide what new UCDs should get added? Perhaps not. But everybody concerned can try to come up with a list of what is currently missiing, and what will probably be needed in the next 10 years (and later on there will be surprises of course), and simply add the resulting list. And keep a mechanism for adding more.
Should the list be changed from a hierarchy to something else? That should not really matter much. (Broadly speaking - software should be able to handle that).
Should the UCDs serve as an ontology for astronomy? I beileve that in a
sense it is. However, I also believe that anything that is related to
astronomy (e.g. astronomers, instruments, observatories etc.) should also
be part of that ontology and could be represented by UCDs in the hierarchy
(adding a few hundred more terms). These can then be used in many other
tables (derived data products). So a growing branch could stay open there.
More later.
Cheers and wishing you all a good time,
-ashish
PS: Does some one plan to email session summaries to the group?
Ashish Mahabal, Caltech Astronomy, Pasadena, CA 91125 http://www.astro.caltech.edu/~aam aam@astro.caltech.edu
Lawyer: How far apart were the vehicles at the time of the collision? Received on 2003-05-12Z05:10:03