We all remember from the Karma days how unreliable user-reported disk problem anecdotes can be. The head unload issue is reported by smartctl and so is almost certainly real: http://www.silentpcreview.com/forums/viewtopic.php?t=51401 although, equally, the variability of the anecdotes on that thread makes it sound as if some people don't get affected; perhaps it's improved in later firmware versions.

Peter