>His theory no longer does hold true. His 1996 paper is available at >http://www.cs.auckland.ac.nz/~pgut001/pubs/secure_del.html, targeting >MFM and RLL disk technology, where a typical 5.25" disk held 20-80MB >(yes, MEGAbyte, not GB). His recommendations were based on old magnetic >disc technology where each bit was represented by the magnetical >orientation on the platter (north=1, south=0). After that came other >technologies, where bits are defined by changes of the magnetic field >even down to probabilistic field measurements - which allowes tighter >data packing but rendered the base of his recommendations useless. Overwriting the data a couple of times seems to be relatively effective for modern disks. In my opninion, the best way to destroy your data is to keep it encrypted and then destroy the keys. Recovery will succeed in only recovering a fraction of the data; but with an unknown key and a typical encryption algorithm where a single bit changed in the input changes all bits in the output, recovery will need to find quite a bit more than the 1 bit per byte required to recover typical text. >OTOH I have seen one company with a *really* thorough disk & tape >cleaning technique: > > 1. writing zeroes all over > 2. low level format > 3. shredding the disc drive into small pieces > 4. magnet treatment of the scrap metal > 5. burning in their own waste incinerating plant Note that many of the people using such technologies do this to prevent against future breakthrouhgs in recovery technology. So it's required only for data with a long half-life. Casper
