Re: TASS Database, versions > 1.0

[aah@nofs.navy.mil]

Chris has offered another modification to the TASS database procedure;
that is, for the local sites to do some of the processing before sending
their data to MR.  As Mike G. mentioned, I doubt all of the sites would
be willing to host Linux/Postgres on their computer systems, so at minimum
this would entail the creation of a DOS/Win95 program to do the matching.
  Chris wrote,
>I am finding that it takes a very long time to do astrometric
>matching when your lists are numbered in the millions of rows.
  This is true, but in general the matching is done on a night-by-night-by-site
basis, and so the 'millions of rows' is averaged over the entire year.
The extra effort over just importing the night's data into the database
without the match should not be that much.
  One of the obvious problems was mentioned later, namely the 'extension'
of the tass catalog.  This is similar to the library checkout procedures
required when several programmers are developing code, as you need to
ensure that the tass catalog is only modified by one site at a time.
  The easiest, most straightforward approach is the one we've already
committed to.  In the ideal world, the best approach would be for each
site to have its own database and the central site just interrogate when
it needs data.  That assumes all sites are running Linux, are nodes on
Internet, are up all of the time, etc.  That is not possible today.
Arne