Tolerances and complex problems

One of my favourite quotations gets a regular airing . . .  indeed much more regular than my blog posts have been recently.

"For every complex problem there is an answer that is clear, simple, and wrong."

H. L. Mencken  (1880 to 1956)

Almost everyone laughs at that.  It sounds both incisive and frivolous at the same time.  I expect most people then dismiss it from their minds and would probably be delighted if they hear it again in a year or two without getting any more wisdom from it.

For me, after enjoying it and regularly using it while reviewing technical designs in a complex engineering environment it has taken on a extra layer of meaning.  It should be the motto of all practical engineers.  What most people don't realise is that the art of engineering is to make things that might be a little bit wrong but to ensure that they are right enough to function correctly.

Engineers call this 'tolerancing'.  Proper choice of tolerances in any design is the art of balancing performance against cost.  If you get the design too right you might not be able to afford to build it, or might not have enough time.

This is well illustrated by the old story about how to tell the difference between a scientist and an engineer. 

Stand them in opposing corners of a room, with an attractive member of the opposite sex naked (and presumably rather tolerant and open-minded) in the middle.  Tell them to approach that person by successively halving the distance between themselves and the object of their desire. Then wait.

The first one to leave the room will be the scientist, who has worked out that, in this way, you will never really get there. The engineer, however, has figured out that you will get close enough - for all practical purposes.