Wouldn't it be better to allow anything that works ? The way we really engineer things ? Its like when I go climbing on an inside climbing wall, where we set up purely hypothetical routes - only yellow holds for hands, only friction for feet, or somesuch. Then when you get on real rock, you have a nasty habit of limiting your options to only certain positions there too, when in reality, on the real thing, what matters is climbing the route...

So if a problem can be solved BETTER with a different variant, or even in 'C' instead of asm, then it should be permitted. BETTER though has to be proved, with a statement of execution time or code-size or even algorithmic beauty.

Steve