Recursion challenge

[LooseCaboose]

LooseCaboose, I think we're deadlocked here. You like you "general purpose" algorithm, and I like mine.

No, I like both for different reasons. The decision of which of the two algorithms to use in practice is dependant on the situation at hand. This is why algorithms such as insertion sort are still used in commercial software today even though other sorting algorithms such as quicksort, mergesort and heapsort are much faster; because it works better for some tasks.

I'll agree that very few (if any) modern machines have word lengths that aren't a power of 2 in length, however a machines hardware word length isn't the only one you would use. For example the RS232 protocol uses 10bit words (8 bit data, 1 stop, 1 start bit) and Huffman codes use variable word lengths. Not so relevant in terms of Gray code generation, just pointing out that irregular word lengths are still very much in use. That said, in about 99% of real world cases, your algorithm would be the much better choice.