Peter Gutmans's Secure Deletion of Data from Magnetic and Solid-State Memory , NIST Magnetic Efforts Page , Schematic of a Magnetic Field Microscope.
IBM Storage division offers a nice tools page. Quantum offers the best SMART explanation. Microsoft and SMART in Windows 2000 Quantum DPS, IBM DFT, Western Digital Data Lifeguard. Storage Review - hard disk reviews.
A primer on magnetic storage from Washington University.
There has always been a lot of discussions as to wether it was possible or not to recover erased data directly from the platters of crashed hard drives. Datarecovery companies have often implied that they were able to achieve such a feat. The DoD recommendations for secure wiping have given substance to the rumour. This article by Science News seems to demonstrate that while some evidence gathering is possible, we are still far from being able to recover data from hard drive tracks. Three problems must be solved to achieve that recovery
Then of course, the eventually recovered data would probably be totally worthless. No one will go through the above cycle to recover business data - this is just not cost effective. If, for the sake of reasonning we admit you just have, would you run critical processes on potentially flawed data ? The same holds for forensic analysis : except in extremely high profile cases such as the Watergates tapes, that could act as a catalyst, an incentive to drive the technology forward, one cannot imagine that courts will be allowed to spend that many resources. And if they did, what would be the values of the evidence recovered ? Small corruptions can completely change the meaning of messages.... Finally, if a governmental agency had such an ability, do you think they would let us know ?
(*) ghost tracks can probably be recovered from floppies where tracks are roughly one tenth of a milimeter wide and magnetic fields, compare that to the 20 Gigabit per square inch common in Y2K's hard drives...