🚀 go-pugleaf

Dave Eastabrook wrote:
> 
[snip]
> >>"Compaq received one of the computers on which the Effect was observed.
> >>Compaq has not been able to reproduce the Effect either through
> >>automated or manual tests.  It did discover a "mis-matched" BIOS/RTC
> >>where the BIOS does not honor the UIP bit and the RTC is not buffered."
> >
> >As they found nothing manually either, nothing is proved about the
> >relative merits of manual and automatic.
> 
> Presuming the automated test was run enough times, then if properly put
> together it *would* have reproduced the problem, since as they say, the
> BIOS did not honour the UIP bit and the RTC was unbuffered  (so the
> contents of the registers would be corrupted during update).  At least
> with the manual test there's some chance.  A proper automatic test would
> of course be infinitely better than fumbling around with a power button.
> 
> Unfortunately the automated test used was obviously crap, not to put too
> fine a point on it.  Unless there's some other explanation?
> 

I've observed automated power-fail testing and one of the items
involved was to not make the loop too robust.  That is, you want
the delays to be random.  Perhaps whatever was put together was
not random enough like it would be when doing a manual test.
Perhaps there is an additional unknown factor that also is important
to create the problem.  Ex:  Clean power?  I'd bet on timing though.

-- 
Are you ready for year MM?
The Mother of all Messes.
239 Days to go before 'Ignorance is bliss' is obsolete.
news:comp.software.year-2000  Come for the signal, stay for the noise.