Vast amount = still limited. Octane uses single precision floating point for its calculations, so the accuracy is about 6 to 7 decimal places.GeoPappas wrote: Sorry, but I wholeheartedly disagree with this statement. The majority of computers and programming languages are capable of vast amounts of precision.
The reason that floating point operations can sometimes result in "almost exact" results is that floating point arithmetic has an inherent rounding error, that comes from it's implementation. So the original problem is not a bug, just how real numbers representation in computers work. Here's a long & techy article about floating points (the "rounding error" paragraph at the top and following, explains this more in detail): LINK
EDIT: Of course most languages have some high-level tools to represent real numbers of arbitrary precision (like BigDecimal class in Java or decimal module in Python), but doing calculations with such datatypes is (much) slower than using the "low-level" primitive float datatype. A margin of error to 7 decimal places is enough to do nice renderings in Octane.
