[mythtv-users] Fwd: Further Notice of Seagate Hard Drive Class Action and Proposed Settlement

[Manuel McLure]

On Fri, Mar 12, 2010 at 2:01 PM, Ben Kamen <bkamen at benjammin.net> wrote:
> With Intel/IBM gear, I always kind of scratched my head because my job was
> to repair intel based PC's and they weren't any more reliable because of
> that 9th bit.

The point of parity on RAM is not necessarily to correct errors
(although ECC can do that in _some_ cases) - it's to detect memory
errors before they can cause extensive data corruption. The idea is
that once you get a bad bit, there's no telling what consequences it
could have. For example, a bad bit in a pointer address could cause
data to be written to the wrong part of memory, clobbering data that
then could get written to external storage. A bad bit in the counter
for a loop could cause you to write 66536 bytes instead of 1000 bytes.
That one bad bit could cause thousands of bytes of data to be
corrupted. With simple one byte parity you stop the processor as soon
as the bad bit is detected, and the corruption has little chance to
spread. This actually may make a machine with bad RAM crash more often
but greatly reduces the chance of extensive data corruption.

On more modern systems, multiple bytes (usually 8) and their parity
bits are combined - this is what's known as ECC. The ECC can detect
and correct single bit errors, and can detect two bit errors but not
correct them. In that case it will crash the system exactly in the
same way as single byte parity does.
-- 
Manuel A. McLure WW1FA <manuel at mclure.org> <http://www.mclure.org>
...for in Ulthar, according to an ancient and significant law,
no man may kill a cat.                       -- H.P. Lovecraft