Errors

["Thomas F. Droege" <droege@fastmail.fm>]

Michael Richmond has changed the processing code so that errors are more
meaningful.  For the new data they certainly look more realistic to me.

Unfortunately, changinging the code does not work back through a time
warp to change the data which has already been processed. I think I will
never go back and process the old images, but if one of you wants to
give it a try there ae about 3000 CDs available.   The new data with the
new errors in now available on mike.  There are four months of data on
mike as tp01, tp02, tp03, and tp04 for   May, June, July, and August of
2005.  September will arrive shortly (What has been taken so far is in
current.)

The data is on mike at /home/tom/current/tpxx   current contains the
current month files.

All the 2005 data has been taken as fixed tiles. Well, the pointing is
not great but is getting better over time.  This data has been sorted
into directories where /home/tom/tom2/ra/dec contains the tiles taken at
ra,dec.  In the directory at the end of the string you will find all the
cal files for that tile.  They have names ra_dec_tttt.ttt where tttt.ttt
is the JD -254000 or some such.  

I have sorted tiles to the nearest degree.  The intention is to go back
and process the old data and make just one big directory of tiles.  The
reason for doing this is that the data has less scatter if we only
compare data with a similar telescope pointing.  Particularly bad is
data in the overlap regions.  Here there is a big offset between data
taken which seems to be consistent.  That is data taken at the top of
one tile will be consistently a different value from data taken at the
bottom of the adjacent tile.  

In /home/tom/current/tile_work you will find the file catalog.  This
file contains all the tiles cataloged so far as:
ra, dec, number of images.  You can just search this file to find, for
example, all the tiles that have been measured 10 or more times. 

Hope this helps settle the confusion.

Tom Droege
-- 
  Thomas F. Droege
  droege@fastmail.fm