I Been Working on the Pipeline

[Thomas Droege <tdroege2@earthlink.net>]

Just to let you all know that automated sky photometery is not all fun and
games I will tell you what "got" me today.

Some things "get" me that you experts would probably avoid.

The TOM2 pipeline failed to produce any I measurements last night.  Looking
at the data it appeared that the sky was brighter than usual but the images
looked flat so I thought I would change the line in sky.param that
determines the level of sky brightness that is accepted.  Mostly I want to
set levels in sky.param and stick with them forever.  However, I am trying
to get measurements for a star for Mike Sallman and I have lost a bunch of
days from TOM2 recently, probably because the limit was set too close.  If
I get results, I may throw them away anyway to make the data consistent. 
But I will send the star to Mike with a caution.  

The sky.param line was:

max_sky { V 4000 }  { I 9000 }

I changed the line to:

max_sky { V 4000 }  { I 12000 }  # I was 9000

The pipeline failed to run, failing on the sky processing.  It would run if
do sky was set to 0 so it must be a problem with the values in max_sky. 
After trying all kinds of values in max_sky, and several hours later I
rewrote the line as:

max_sky { V 4000 }  { I 12000 }  
# I was 9000

and it ran.

OK, it looks like the # is ignored in a line where data is expected and who
knows what is used for the parameter.

Sigh!  Just to let you all know that I have to work at this.  I am not
complaining, and I certainly don't want others to change coding style to
make things easier form me.   It is my job to learn.  Something like this
caught me once before or it would have taken much longer. 

Meanwhile one of the dual computers has died that runs the pipeline.  So I
will have to fix it and also need to try to rescue the data for TOM2l and
TOM3e for the month of July.  Sigh!  I have the images so at worst I will
have to run them again to rescue the data.  Hopefully is is something like
a bad fan and the computer will recover. 

Judging from the comments on the AAVSO list, it must be nice to just go out
into the night air and look at stars and estimate their brightness.  It is
not so easy for automated systems to do away with visual observations. 
Still I will get a million or so measurements when I straighten out last
night's data.  A few days of work for me, I suspect.  The trick is to fix
the last batch of data before another clear night comes along.

Tom Droege