The following needs to be documented in Help:
1) Precision of integers and double floating point variables.
2) Built-in functions for floating point variables:
exp, ln, log, sin, cos, sqrt, random (uniform), fabs, pow, ceil, floor
(some more candidates: tan, asin, acos, atan, sinh, cosh, tanh, asinh, acosh, atanh, log2, exp2, min, max)
3) expression types in weights of edges and exponential rates on locations.
4) conversion rules between double and integer.