>> hi, for me that sounds as the disc said 'huh, i wasn't able to 
>> decipher
>> what was on my platter, so i read again, proved it and wrote +1 to the
>> counter'...? jeeeeez.
>
> The errors are corrected in the read-buffer (cache) without any 
> performance
> loss.

that is one-bit errors corrected via the reed-solomon algorithm?

but i still wonder why wd1 had 6 millions of correctable errors while 
wd0 had 'only' 25 thousands...

> Yet this is only a fraction of the magic modern drives use
> to acheive current data densities ...

yes, GMR is nearly out-of-date ;)

regards,

timo