How does C2 calculate number, precision-wise?

Just curious here, are the number variables we are using double, float, or something else?

How many digits of number can we go for, until we start to have a little round off error, precision error, etc?

Now Ashley can correct me if I'm wrong, but I believe C2 just uses JavaScript's logic and datatypes. Now JS is an untyped, or dynamically typed language. That means that you generally don't declare what type a variable is going to be, it's taken care of behind the scenes. This makes predicting round of errors and precision errors a nightmare because it's all guesswork on what type JS is choosing at that point in time. As a general rule, JS will pick the most suitable type, so:

"asdasd" = string

65161 = int

65651651684651654165168416517165141651 = long

1/3 = double

1/4 = float

Construct 2 leaves number calculations to the Javascript engine, and Javascript is standardising as only supporting double precision floats for numbers.