On Sat, 10 Jan 2004 16:38:42 +0000, Terry Pratchett
<tpratchett@unseen.demon.co.uk> wrote:

Nevertheless, I will say 'nothing', on the basis outlined below.  And 
while I'm well aware that it isn't the computer that puts the coffee cup 
on the CD tray, Windows is quite capable of throwing the occasional 
spanner  into its own works.

That is still the programmer causing the problem same as the Sears
incident.

The last 'short' unit (6hrs.03 minutes) was given to the machine by 
SetiQueue  at 03.20 on Jan 1.  I could be wrong, but that leads me to 
believe that nothing magically happened at midnight, but possibly 
happened while that unit  was being run.  The next unit, given to the 
machine at 9.24, took more that 15 hours to run.

Okay the next thing to check are the angles, more specifically the RA
angle. Check in your SetiQueue and see if you aren't receiving lots of
very low or very high angles. An average is around .423 or
thereabouts. Much difference than that neighborhood and the computer
can take a very long time.
You could also be working your way thru workunits that have lots of
Spikes, Triplets, Gaussians, or Pulses. I believe that Spikes and
Pulses will extend the time needed for crunching.
I believe that you said that you deleted Seti from your computer and
redownloaded it and reinstalled it? If not I would do that if neither
of the above seem to pan out. 
OOOOH....have you defragmented the drive lately? If yes than that
could have moved Seti to new places, if not that could be part of the
problem especially if the hard drive is almost full.