A computer will typically represent a number using a fixed, finite amount of memory. This can be problematic when precision is important because information is lost when a number possesses more information than the space allocated for its representation can possibly store. Fortunately, Wolfram|Alpha has the power to do computations in arbitrary precision, eliminating the cumulative error that arises when fixed point, floating point or conventional integer representations are used.
Utilize integer representations that are free of range boundaries.
Do exact arithmetic with large numbers:
Utilize decimal representations that have arbitrary range and precision.