Just bumping this incredibly important article. Most, if not all visitors to this site have encountered broken complex systems, and are even in possession of a variety of failed fixes or ever-broken machinery or tools. What's interesting here is the identification of non-intuitive problem-solving techniques, specifically these points:

-Identify people who run the system and invite them to be a part of the trouble shooting team.
-Could it be that the operator changed and the root cause is a person and not the system?
-Until proven wrong, assume there is more than one source of failures.
-If experts exist outside of the team, ask them to attend periodic reviews. Often the simple act of preparing for these reviews will shed light on the problem.
-Beyond the technical effort of finding the root causes of the failures is the political side.