Okay, with g(2, b, c) = c^b the definition makes sense. But it's a little weird, it reverses b and c from their usual order. Every other version I've seen has the normal order. I think it would be better to either have
g(a, b, c) = a ^{c-1} b
or
g(a, b, c) = b ^{a-1} c
The former is the one on Wikipedia, if that matters. Deedlit11 (talk) 07:33, March 13, 2013 (UTC)