Actually, guttmann has published a paper which denotes that 35 passes isn't a magic bullet to solving that dilemma, In fact he postulates that you need basically a unique approach to nuking information on your unique harddrive model. I also like his theories about RAM memory "burn in" which reveals that your harddrive may not be the only insecure media that you're using.
However, the degausser approach also sounds like a similar idea, so I don't know which is the more secure alternative here. However, if you're looking to using your harddrive again, the degausser approach sounds alot more nicer to me since it deals with the disk as a whole, and doesn't need as many overwrites to blank the disk.
This text makes my post look better.