Subject: Re: SAH/BOINC crash and burn
From: f/f george
Date: 22/09/2004, 04:29
Newsgroups: alt.sci.seti

On Tue, 21 Sep 2004 19:35:35 -0400, "Dave Gower"
<davegow.removethis@magma.ca> wrote:

As I mentioned yesterday, my SAH/BOINC was not working properly and wanted 
me to reset. Once I found out how to do that, all hell broke loose. First, 
it tried to bring down a bunch of data sets, and every single one failed due 
to some sort of missing file. Then it tried to dial when I didn't want it to 
(see below). When I tried to cancel the dial, it hijacked my computer and 
refused to let me cancel the dial (I'm not kidding). Then when I managed to 
close the connection, it erased my shortcut.

That's not all. My computer room is across the hall from the bedroom. Twice 
last night I was woken up by the computer dialling. It was not finished its 
assignments, and my preferences were set a) to once a day and b) ask me 
first. So both times, and then the time I referred to above, it was ignoring 
my preferences.

I uninstalled SETI/BOINC completely. I'm really sorry gang because I still 
think SETI is the most important intellectual activity going on in the world 
today, but this is just too much. SAH/BOINC is obviously not ready for 
release. If this were a commercial operation, some people would be fired. 
SAH/BOINC is crap. Rubbish. An insult to the hundreds of thousands of people 
who really want to participate. A real betrayal of science. 

It probably kept trying to connect because it wasn't updating itself
to the new settings yet. It only updates the local computer during the
next connection or if you click "update" manually.
As for the connection and crashing issues I think the program was
midstream when you tried to make it stop and since Windows HATES that
they conspired and Boinc lost the battle.
As for quitting I would do as others have suggested, go back to
Classic for awhile and crunch unitl they make you stop. You WILL be
contributing, the units being worked on now are "interesting" ones
from before and need to be crunched a few hundred times each to verify
whatever results are being sent in.