In the wake of Google's eye-opening report on hard drive failure comes this follow-up from Ars Technica, which states rather flat-out that the curious problem of drives that fail without warning isn't going to get better any time soon.

The problem? No one really knows why drives fail, and while certain drives failures can be traced to SMART errors, other variables have been elusive in pinpointing what exactly makes hard drives crash. Even within SMART errors, only a small subset (four, in fact) were found to be of much importance in determining whether a drive was headed south.

Additional information has surfaced, thanks to more expert testimony and another large-scale study, this one from Carnegie Mellon. The results are unfortunately contradictory: The CM study found no special tendency for drives to fail early in their lives, while drives over five years old were found to be 30 times more likely to fail than usual. 

But we can debate drive failure causes day and night; unfortunately no one cause (SMART, age, heat) can pinpoint any drive's impending doom with any degree of reliability. The real issue, according to the story, is that no one really much cares about building drives that never crash. Although the piece rightly notes that hard drives have become a kind of disposable commodity (with a two-year lifespan), it doesn't mention that old drives quickly become so limited in capacity relative to their newer brethren that no one wants them, whether they're working or not. So older drives eventually find themselves upgraded, failure or no. Acceptable? Not really. But it's a reality.