[Jim Sims]

Moving out a little more globally (away from a focus on structural failure), failure can be defined as not meeting specifications.

A machine that never breaks or corrodes could be considered a failure if it is too heavy, too expensive, or too inefficient is the use of energy.

I propose that all failures of man made artifacts are ultimately traceable to man and his lack of knowledge of system boundary conditions, material properties, and all possible permutations of conditions in complex systems (control software failures).

Failure to compose or apply the proper specifications is the root of many failures.

Misuse or abuse of a system is merely a matter of definition and expectations.
Most would consider wear out of an internal combustion engine in one or two thousand cycles a failure. However, this is the typical lifetime of a high performance artillery piece barrel which is merely a highly loaded internal combustion engine.