TASS Errors

[Thomas Droege <tdroege2@earthlink.net>]

A little more about errors.  The curve of S&T note 10 is the error
distribution for the whole data set, all points, all nights, variable stars
included, everything.  

One would think it would improve if "obviously bad" data points were
removed such as the Michael Sallman "flagged" set.  It does not.  I have
made many cuts and even selected out the very best data.  The error curve
still looks the same.  This has led me to believe that it is something else
that generates more error than one would hope for.  Michael Richmond has
been investigating some posibilities, I have been looking at others. 
Andrew Bennett has also looked in the past.  

I know that if I track a star, the error is much lower.  Almost a factor of
10.  This has let me to suspect the reference catalog.  We are forced to
use reference stars with large errors because we don't have a catalog that
is better.  Horrers you say?  Yes Virginia and the rest of you, even the
well publicized catalogs have errors in them.   Big errors.  Possibly we
can eventually produce a better catalog but we don't have one yet.  

Hmmm!  It looks like I am having one of those days.  The spell checker just
quit working.  gxxx is a good word now though it would have been an error a
few minutes ago.  Sigh!  I have looked in vain for where the switch has
been turnde off.  

Tom Droege


Tom  Droege Jennifer Malpass
tdroege2@earthlink.net