Re: check to see when TASS cameras were taking data

[Chris Albertson <chrisja@jps.net>]

Stupendous Man wrote:
> 
>   There's a new item on the TASS "Access the database" page
> 
>          http://a188-l004.rit.edu/tass/dbms_access.html
> 
>   If you choose "Check to see when TASS cameras were active",
> 
>          http://a188-l004.rit.edu/tass/test_timeobs.html
> 
> then you can generate either a plot (by default), or an ASCII list,
> or both, showing when each of the TASS cameras was running during
> some specified time period.  This may be handy for people who
> are interested in some specific event.
> 
>   Please send me comments/suggestions -- I'd like to improve it
> so that it meets the needs of the people.
> 
>   One longer-range goal I'd like to implement is a means to
> use this list of cameras and times to figure out the _area on the sky_
> that was being monitored as a function of time.  That would be very
> useful for people interested in checking on GRBs, supernovae, etc.

One of the neat features of Postgres is that it has in addition
to the "standard" SQL data types a set of "geometric" types.  Things
like point, box, pollygons, lines and so on.  It has boolian operators
like "point in box" "box overlaps box", and so on.

It also has the ability to accept user defined types and operators.
The example in the book shows how to define a type called "complex"
and a set of operators "+", "-" "*"... over complex numbers. 

I have been thinking we could use a set of functions that operate
on RA and DEC.  We can store the frame boundries as a BOX type
then query for "the set of boxes containing a given point".
That should give you the set of frames where an object _should_
have been seen.

I have taken some time out to figure out how to add user defined
functions.  My first goal is a Standard Deviation function.  Once
this is done there will be no need o write special purpose Perl or
C programs to compute Sigmas.  Next comes a function to compute distance
between two stars with RA and DEC in mas units.  So far it does not look
hard to do.  The code all fits on one screen.


> 
>   But that will take me a day or two, sometime in the future.

The "source file" table contains the bouning box for each frame
of data.  You could search there.  It shouuld be faster as there
are ~2000X less frames than stars.

Once we define some operators over RA and Dec (they need to know about
wraps at the equator and poles) it will be easy
> 
>                                          Michael Richmond
> 
> 

-- 
   --Chris Albertson             home: chrisja@jps.net        
     Redondo Beach, California   work: chris@topdog.logicon.com