More on algorithms that reveal properties of floating point arithmetic units
In the interests of producing portable mathematical software, it is highly desirable for a program to be able directly to obtain fundamental properties of the environment in which it is to run. The installer would then not be obliged to change appropriate magic constants in the source code, and the user would not have to provide information he may very well not understand. Until the standard definitions of programming languages are changed to require builtin functions that provide this information [1, 3], we will have to resort to writing routines that discover it.