Disaster, sort of.

[Tass Mailing List <tass@mail.alembic.net>]

Date: Tue, 17 Jul 2007 12:38:11 -0500
From: Thomas F. Droege <droege@fastmail.fm>
To: tass@tass-survey.org
Subject: Disaster, sort of.

Hi Everyone,

I remind you all of what I was trying to do as I set up tass.

1) Do a bit of real science
2) Expose the real scientific process with all it's warts and problems.
3) Entertain my audience

We now have a prime example of 2).  Stuff happens, and it has happened
to me.  I have worked on enough large experiments at Fermilab to know
that the below type of problem is not unusual. I assure you that it
happens even to Nobel prize winners.  For those in the know about such
things, I designed the apparatus that discovered the "Split A2".  Not
really the fault of my stuff though.  Good science involves "confessing
to the problem" and then fixing it.  We are in the process of the
confession phase, and I am starting the fix.

Last month I discovered that when I set up the processing files in May
of 2005, I made the mistake of setting my site longitude to -88,33
degrees if longitude instead of the +88.33 degrees expected by Michael's
program.  Sigh!  The instructions in the .param file were quite clear.
Still, -88 is what one gets if one goes to a map service to check on my
location.  This resulted in big errors in zero point for the processed
magnitudes.  Well, not so big as to attract attention until I noticed a
few mag 6 changes and started looking into the cause.  The really bad
thing is that there are also small errors.  So it is not just the
program trying to look through the solid earth at the measured star.

This means that all the data since May of 2005 for tom1 has to be
reprocessed.  Somewhat later for tom2 and tom3.  Fortunately, the
patches catalog uses very little of the "bad" data so it will only need
correction at some later date.

I am setting up to reprocess the data.  Unfortunately, it looks like I
can't make it completely automatic.  I was just "creative" enough in my
file saving to make me have to look at the directories of each DVD to
see how the data was saved.  For example, I know of at least one case
where two days of data were saved on a single DVD.  Seemed like a good
thing at the time since it saved three DVD's, but now I will have to
look at each directory to find that day.

I hope to be able to run several computers at a time to process the
data.  I have a row of 4 set up.  The general scheme will be to load
each with several days of data and then just let them run till the next
day.  I anticipate several hours of loading before the processing can
start.

I plan to start with the tom2 data so that I can get a long time series
of "good" data as quickly as possible.

I think there is nothing others can do besides shout encouragement.
Suggestions as to things to do during reprocessing would be appreciated.

On the good news side, I am still here to do the work.  I really
expected to be dead a year ago.  Some might say that I have been spared
for a reason.  So, off to work.  I will report progress.  I will report
a list of tpxx files that need to be discarded as I set up.

Tom Droege

-- 
  Thomas F. Droege
  droege@fastmail.fm