4. Quantum Leaps in Mechanical Engineering

There's a grinding reality about physical things that computer scientists just don't get: On the human scale, physical things are wholly chained to Newton's Laws of Motion. Computational cycles can improve logarithmically but f shall always and forever equal m*a.

I "invented" a new valve system for the 4-stroke internal combustion engine when I was 16. Took it as far as machining parts to try it out on a Briggs & Stratton rototiller. It was going to make me rich, it was - I'd solved valve float. There were going to be problems with leakage but that was incremental. It was well worth the struggle.

At 18 I caught a segment on the late, lamented Beyond 2000. Some goddamn Frenchman had stolen my idea without ever meeting me. Worse, he'd built prototypes. Worse, he was racing Formula 1. Still worse, he'd stolen the idea from some dude back in 1938.

It answered one question - the design worked. It answered another, too - leaking was a problem. I lucked out, though - the Frenchman turned a large fortune into a small one pursuing a problem that engineers had licked a dozen different ways for half a dozen decades. I spent some time studying vehicle design and stumbled across four other valve designs similar to mine. All of them worked sort of; none of them worked quite as well as the conventional poppet valves that have been the mainstream of Otto Cycle engines for a hundred years now.

The equations are complex, but they are repeatable. The math can be done. If it's bigger than a breadbox and sounds too good to be true, it is.