Doing science with prototype Virtual Observatory tools

[Anita Richards]

Having finally read this in detail, it is very interesting and will be
useful in planning the AVO Jan 2004 demo, especially for other things to
do with the GOODS data.

Some comments:

1. Significance of results.
The tests described involved looking for clustering and correlations in
data set federations.  Where these appear to produce interesting results
we need to know the statistical significance of the fit of the correlation
to the model, either using just the internal dispersion or using published
errors.  I am probably the referee from hell as when I get a paper with a
plot which "clearly shows y is related to x by ..." then either it is
too good to by true and I look for selection effects or it is clear as
mud... including probabilities would cure the second problem.
VOPlot has some basic statistical functions, and I have actually had more
luck than Mark and Bob getting it to do what it claims, a review of these
across packages would be a nice future project (or just get them into
VOPlot?).

2. Topcat - what is 'not too much' Java?  Less than one line?

3. Minor details:
Fig 13.a The Chandra errors seem rather large, they seem to be ~2" in the
catalogue; I have had problems in Aladin with symbols drawn by a filter
not rescaling properly when you zoom.

Inevitable advertising :) adding radio data would add to the science
value, e.g. the last para of 5.4.1, about embedded AGN in starburst
galaxies. In the HDF(N) we found a slight (but probably real) misalignment
between the fainter Chandra and radio sources; moreover there seem to be
more cross-IDs between faint radio sources and soft X-ray sources than
hard.  This would support the idea that the faintest radio sources are
mostly starbursts which might have soft x-ray emission but they might also
have a separate x-ray source in the same galaxy.  In the CDF(S) there are
two frequencies of VLA data so that could give even more information.

4. I have just used Aladin to investigate the proper motion of a star (U
Ori) to see if some MERLIN 2001 water maser observations are centred on
where the star could reasonably expect to end up.  Using Aladin took me
about 3 mins to check what direction the star appeared to be heading in
and what catalogues would be useful to use to find the proper motion
accurately.  Without Aladin the same could be done but much more slowly.

Cheers
a


- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Dr. Anita M. S. Richards, AVO Astronomer
MERLIN/VLBI National Facility, University of Manchester,
Jodrell Bank Observatory, Macclesfield, Cheshire SK11 9DL, U.K.
tel +44 (0)1477 572683 (direct); 571321 (switchboard); 571618 (fax).