Software floating point with type-tagged variable mantissa precision, implemented using a strict pair of Integer and Int scaled alike to decodeFloat. Version 0.4 adds more number-type-agnostic numerical algorithms (sin, cos, tan). Instances of the usual numeric type classes are provided, along with additional operators (with carefully chosen fixities) to coerce, adjust and reify precisions. The intention with this library is to be relatively simple but still useful, refer to the documentation for caveats concerning accuracy and assorted ill-behaviour. Usage with ghc(i)-7.0.4 might require -fcontext-stack=100.