Yeah, I'm already using that for a few things, like the squish.
y = n - (n-1)/x {1<x<inf}.
When X is 1, Y is 1, but as X increases, y gets closer and closer to n without ever reaching it. It's good when it needs to always perpetually increase the more you interact with it, but still need to have a hard maximum otherwise it'll look strange.
As for using that to handle floating points, it's not possible. At least, not the way I'm handling it. 1.7e308 is, give or take, the maximum value for a variable I use, but the problem with that is this means that I can never, in any function, ever exceed that number. I can't exceed it, detect that I've exceeded it, then divide it by some value, because the exact moment it exceeds that value, it becomes infinity and there's no reclaiming my original number anymore.
I'll probably end up just having it detect if the value is infinity, in which case it'll display the largest fox it can without letting it grow anymore.