Re:Inexact floating point calculations... (Score 1)

Mathematica uses symbolics (sqrt(2) vs 1.4142...) until you actually ask for a numerical answer. Mathematica can do arbitrary precision floating point (via N[expression,sigdig]) when you want a number "made" out of your symbolic result. (This presumably takes an absurd amount of time if your symbolic result is complicated.)

--Griffy, Server Janitor, Math@Univ of MD