Tim wrote:
On Mon, 2008-04-07 at 12:53 -0400, Gene Heskett wrote:
I had an alcoholic uncle who, during WW-II, worked as a postal delivery type in
downtown Des Moines, and he usually had a radio in the bottom of his mailbag,
either going back to a customer after repairs, or given to him to be repaired
while he walked his route. I think I was about 8, maybe 9 when I asked him
as he was changing the filter capacitor in one of those 'all american 5
tubers', what was actually wrong with the part he was taking out, and he
couldn't tell me! He was going by a recipe on the inside of the cupboard
door that said to change them if the radio had a hum.
I remember doing work experience at a local TV repair shop while I was
studying electronics. They had this set full of gremlims that had been
plaguing them for weeks. Their best tech had spent ages repairing one
fault after another, but it was still coming up with more things needing
fixing. After he expressed frustration, one of the other road techs
(the ones that do the house calls) had a bash at it, literally, to
finally get rid of the thing. Poking and prodding boards, rather
roughly, with the plastic end of his screwdriver, finding everything
that was unreliable and needing replacing. The set that had be worked
on, off and on, over weeks, was fixed in about a quarter of an
hour. ;-)
In 1967 or so, when a mainframe CPU was still the size of a phone booth,
we were running a GE 605, which was a later enhanced mil-spec version of
the 635, and like all computers made of plug-in boards with components
on them, it was pretty unreliable. Then we got a new FE who took the
machine down, reseated each little board with a thumb and tap with a
plastic mallet, and started diagnostics. As each bad board was found is
was tossed down the computer room floor and replaced. Then doagnostics
were run and each board got hit with a hair drier (same toss on the pile
of dead boards), followed by the can of freon. He didn't even read the
console, if it clattered the board was history.
After that we ran for 35 days, and a counter in the OS overflowed and
the system crashed. There were about several hundred of these in use,
they had been out for five years or so, and no one had ever hit that
overflow before! What does that say about reliability...
In 1996 I booted a 386sx-16 on Linux 2.0.13, using computer show parts,
and ran until Y2k. Unless you were around in the bad old days where you
reseated memory chips with every boot of your "home computer" and
counted crashes per month on mainframes, you can't really get a feel for
how good we have it now.
The machine was called glacial, because of performance rather than
because it was cool, but it served DNS for 4+years. There was a counter
which overflowed in Linux at that time, but it never bit me.
--
Bill Davidsen <davidsen@xxxxxxx>
"We have more to fear from the bungling of the incompetent than from
the machinations of the wicked." - from Slashdot
