I have to disagree with Dennis' proposal. I want to see some mathematical evidence that a defrag operation shortens disk life. Sure it sounds possible, but can anyone produce produce these numbers from a scientific study or from a hardware manufacturer? I have to ask myself, self, that's what I call myself, why is a DJ focused web site the first place that I have heard of this? I have been in IT for 20+ years, never heard anything about this and I hear about it on a DJ web site? Hmm, I guess it is the internet, so it must be true...
We used to run nightly defrag operations on server hard drives back when had physical drives. Now with massive SAN & NAS systems that is basically impossible, but we still defrag local server boot drives and we highly recommend that users defragment their local system drives. Defragmenting not only speeds up access, but by keeping files in contiguous blocks on the disk it also enhance any chance of data recovery from corruption.
From what I have read, and experienced, heat is the number one killer of hard drives. Physical abuse would be a close number 2, especially with laptops. Even carefully moving a computer with the drive spinning is asking for physical contact between the heads and the platters. We tell users that they should use the suspend/resume feature when walking from their office to a conference room.
Now back to the original question - 80% is a good rule of thumb for considering a drive as being full, espcailly when you have a single drive acting as the OS, temp, swap & data device. Crossing that boundary with almost any operating system and there is usually a significant noticeable drop in performance. If it is a dedicated data drive with static data you can go all the way to full without a performance penalty.
Scott