Greetings and Inflammations!

>der mouse proclaimed:

>> I *would* be curious, though, to hear if others *have* seen memory
>> problems that truly were "defects" and not just "poorly installed"
>> parts, etc.
>
>It may not be quite what you want, but in my collection of RAM, I have
>three SIMMs that I keep very carefully separate from the rest.  They
>are "known bad", and I never put them into a machine except to check
>whether its memory self-test is capable of catching them.
>
>However, I feel sure all were in service at some point, and quite
>possibly failed in service.


Yes, but they *failed*.  Undoubtedly in ways that were caught
by self-test with or without parity!  Haing parity just gave
you a bunch of traps while they were running.  Rerunning the
memory test would undoubtedly have also turned this up.

>I've also had Sun-3/60s start to crash randomly in ways that I
>attribute to overheating of the RAM.  (Replacing the RAM and putting
>them in enclosures with good cooling made the crashes go away.)


That falls into my "you have *other* problems" category.
Either the enclosures were designed improperly, you stuffed
too many watts into them or the fan failed, etc.

E.g., years ago, 7200 RPM drives were considered "flakey".
In reality, it was their use in poorly ventilated cases
that was the problem.  Had they been installed by "professionals"
(ha!  what's *that*??  :>) at the "factory", this probably wouldn't
have been an issue (especially if those "professionals" had to
warranty their work!).

So, the drive/memory/whatever isnt the real problem -- just
a place that problems manifest themselves!  ("Gee, I'm having
lots of memory errors on my 3GHz overclocked Pentium 2..."
"Ah, you need to reinstall Windows..."  :>)

--don