What's the best way to implement and model integers of unbound size? should I use fd or just just church encode them? I'm building something simple with only lists and numbers. The only numerical operation it should support is increment