5 Dec
2006
5 Dec
'06
11:11 a.m.
Tony Duell wrote:
True. That's why I said 'in the same
sense'. It's quite possible for
hardware to have worked perfectly for 20 years, and then to stop working
(and not work agian until it's repaired) because some component has
failed. I think it's very unlikely for software to work fine for 20
years, then crash and not run again because of some latent design bug
(unless said software makes use of the real time/date, of course ;-))
I have a machine that ran Linux for years, but suddenly
started to hang at random times for no known reason.
Windows 98 ran fine, only Linux had a problem.
I replaced the hard drive, installed a new version of Linux,
and things improved for a month. Then it started hanging
during boot. Windows 98 still ran fine.
I am now replacing several capaciters on the motherboard,
with obvious buldges on the top, and Linux again runs fine.
I still have 4 more caps to go, but need to find ones with
the correct form factor (tight fit). Pulled those I used off
another motherboard that had lightning damage.
I don't know why Windows would run, but Linux would not,
except that Linux makes more use of the motherboard features
than windows would.
I have had experience with one software bug that went
unoticed for years. Then one programmer noticed the bug.
Almost immediately, the software failed because of this evry
bug on numerous systems, where it had run for years.
Eerie.