Re: TASS Database, plans and progress.

[Stupendous Man <richmond@a188-l009.rit.edu>]

> I now have (as yet untested) code that splits the observation table up 
> into multiple tables.
> This is required to get around the 2GB per file limit imposed by most
> 32-Bit OSes.  The software first splits the data based by filter then into 
> ranges of JD.  At the current rate of data collection each table can hold
> about one year of data.  This had to be done as the current "everything in
> one table" design would overflow to soon.

  I've worked my way through most (but not all) of the data which
Glenn, Tom, Mike G., and the JHU guys have sent in.  The number
of entries in the "observ" table (the one containing one entry per
measurement of a star) is about 6 million, and there are about 
2 million distinct entries in the "tass_cat" table.  The largest
file in the Postgres data area is 606 MB.

  I agree - we'd run out of room pretty quickly if we couldn't
create files larger than 2 GB.

 

> 4) ability to store and output calibrated photometry.  I'll
>    supply a dummy procedure that computes a constant color
>    transform.  Someone else (Arne?) can modify it so that it
>    actually computes something  (not started yet)

  Sigh.  I keep hoping for a mythical "day of freedom" to work on
this calibration step  :-(


> 2)  re. #4 above:  While it is normally not done this way,  I
>     assume a color transformation can be written as
> 
>               V = a*v + b*r + c*i
>               R = d*v + e*r + f*i
>               I = g*v + h*r + i*i

  Yup, this can be done.  It will confuse the heck out of traditional
astronomers, but they deserve it :-)


> 3) The dummy calibration routine will accept the following as input
>      a) An array containing all of the Landolt data.
>      b) An array containing TASS observations of Landolt stars by one
>         site over a defined period of JD (nominally one night)

  I hope that this may be modified with little trouble so that
the "standard stars" may be any set chosen by the user -- specifically,
a set of stars from the Tycho catalog.  If the routine just asks
for 

       a) an array of "standard star" data
       b) an array of TASS observations of this stars

then the user can choose Landolt, or Tycho, or some other set
of data.  I would guess that this is what you had in mind, anyway.


>    Question: should anything else be computed, like for example a
>    "degree of fit" or error number?  and if so does this need to be
>    stored with the matrix? and then how is it later used?                   

  It would be convenient to calculate the scatter between the
"standard catalog" magnitudes and the transformed TASS magnitudes,
after the fit has been made; this gives some indication of the quality
of the calibration (if the scatter afterwards is +/- 1 mag, then
you know that something is fishy).  It's not really necessary to
compute it as a part of the calibration process, because someone can
always take the calibrated TASS data for the "standard stars"
and compare it to the "standard catalog" data, afterwards ... but
it would be nice to do it during the calibration step -- all the
required data is at hand.

  Sounds good, Chris.  Have fun in Texas :-)

-- 
-----                                      
Michael Richmond                   "This is the heart that broke my finger."
mwrsps@rit.edu                     http://p674p06.isc.rit.edu/richmond/