Ack! Hard disk errors! Surely, "this is the stuff that PC nightmares are made
of". 
Fortunately, true errors are rarely encountered when using hard disks. To
help users understand better the rate at which errors will occur with a hard disk,
manufacturers provide anywhere from one to several error rate specifications.
The most common error rate spec is the drive's unrecoverable error rate, which
is usually specified as "<1 in 10N bits", where "N" is
usually between 12 and 15. "Unrecoverable" means that the drive is unable to use
its error-correcting code, retries or other
techniques to recover from the error in reading the disk and thus properly recreate the
data. If "N" is 14, then that means this will occur every 100 trillion
(100,000,000,000,000) bits read from the disk. Not too shabby.
In fact, drives usually have several different error rate specifications; they are just
usually not put into the short data sheets commonly distributed by hard disk
manufacturers. In fairness, unrecoverable errors are the most important ones, but there
are also specifications for recoverable errors, errors recovered after multiple reads and
so on. To find these, you generally need to download the product manual for the drive or
contact the manufacturer's technical support department. For a full explanation of these
various errors and what they mean, see this full
discussion of errors and error recovery.
Error rate specifications are typically used to compare drives. Within the same general
class of drive there are usually relatively few differences between manufacturers. The
biggest difference in error rate can be seen by comparing newer drives to older
ones--newer drives are usually significantly better despite pushing the design
envelope with higher speed and much greater capacity.
Next: Warranty Length
