Errata exists for this version of the document.

This Simple DataType defines a high-precision signed number. It consists of an arbitrary precision integer unscaled value and an integer scale. The scale is the inverse power of ten that is applied to the unscaled value.