The article outlines five software-based alternatives to standard floating point number representation: decimal floating point (base 10), fractions for exact calculations, symbolic computation, interval arithmetic for error tracking, and binary-coded decimal used historically in IBM systems and financial formats.