Seth Goodman <seth_news@yahoo.com> wrote:

Darrel Toepfer wrote:

http://seticlassic.ssl.berkeley.edu/sstatus.html
The data server process is not running
Right now the data server machine is up...

snip

Been offline for around 15 to 20 hours @ best guestimate...

The server has been up and down since yesterday. I just downloaded a 
classic WU, one minute ago, so it's not dead yet!

I've been downloading WU's during this period, although with
some difficulty connecting. However, in addition to the
message at the beginning of each WU that says we need to be
changing to BOINC, I'm now getting a "duplicate" error
following each upload that causes the Seti program to shut
down. This may be due to the "unknown" ErrorCode 0x00002f78
12152 mentioned by / Johan in Message-ID:
<438EF031.24633C9E@telia.com>. 

Most, but not all, of the time the total on the user stats
page is being updated anyway, but the Seti program is not
accepting and reporting this total.

I had hoped to process another 150 or so WU's before the
project shuts down, but the current conditions are making
this difficult. I would like to think this is not
intentional. 

I probably will not go to BOINC for several non-pejorative
reasons. I'm using too much electricity by leaving computers
running. Newer, faster computers tend to be noisy without
special measures and are preferably turned off when not in
use. Newer CRT's go into amber mode and therefore have less
need for screensavers. (Seti Classic was too slow unless the
screensaver was turned off, anyway.) LCD screens don't burn
in (I think) and therefore don't need screensavers at all.
Newer computers boot faster because of faster processors and
Windows Hibernation and therefore provide less incentive for
leaving them on. The fastest Athlon XP's (and perhaps the
Intels as well) are hard to cool, and leaving them on sucks
in more dust, requiring cleaning. Lastly, Win2K and WinXP
are sufficiently stable that one can reasonably expect them
to boot without problems. 

On the other hand, excessive thermal cycling increases
motherboard failure. It also seems to cause the heat sink
compound to gradually work it's way out from between the
CPU-heatsink sandwich. I was told the compound on XP 3200's
should be replaced at least every six months anyway. This is
not a pleasant prospect.  Someone actually suggested the
main reason AMD introduced the Athlon 64 was to give an
excuse to make the die bigger, and therefore easier to cool.
I haven't had a chance to test this out.  I don't really see
64 bit OS'es and Apps lining up, except for BOINC. Then
there is the cycle-soaker argument, which has been going on
at least since the IBM/370 days, if not longer. Something
about quantum mechanics and "keeping the atoms busy", I
think. 

Dave