It was pretty long as it was... Tacking on a code generation section to show why certain things are inefficient (and the associated detour to how the standard library is implemented which would be required reading to understand all the implications) would make the article three times as long. I promised I'd have it done by Monday.
