> rather than solving for success, solve for failure. If you want to design a dust filter that works at 600 Celsius within a steel mill, instead design all the ways you could develop a filter that would fail miserably in those conditions.

That's a strange example. There are infinite ways of not accomplishing a task, and as a non-expert in filter design I have no trouble listing them: make the filter out of butter, make it just a big hole, pay someone to pluck dust out of the air with tweezers... how exactly is this helping?