All hard drive errors boil down to how many failed bits occur on the raw, pre ECC corrected media vs. the calcuated post-ECC return. A hard failure is one that exceeds the span of possible corrections. Most hard block failures should be correctable by sparing of the media block in question. If you get too many non-correctable errors, it is indicative that the electronics or the heads have died... which in practice turns out to be a catalysmic failure where the drive totally fails on a subset of reads (i.e. one surface is no longer accessable).I would think a better way to test a drive would be to perform long reads (data + ECC), programmatically calculating ECC and determining the number of bits in error, and then performing sparing of the problematic tracks (if supported by the command set of the drive - SCSI does this, I don't know if ATA drives allow sparing to occur in call cases.) Of course, a simple 'dd if=/dev/sda of=/dev/null bs=1024k' may be just as effective in the long run...
Arbitration mentions in a contract are a lot like the signs at the go-kart track - they are designed to make people think that they have no legal recourse. To paraphrase a previous poster, judges don't take kindly to those who say they don't have a say in a situation that is placed in front of them. These clauses often get thrown out if a lawyer is involved, from what I have seen.
"The Mets were great in 'sixty eight, The Cards were fine in 'sixty nine, But the Cubs will be heavenly in nineteen and seventy." -- Ernie Banks