Some historical context... IEEE 754 (the current floating point standard in H/W and S/W) arose in 1985, only 6 years before this paper. Before that, floating p

Some historical context... IEEE 754 (the current floating point standard in H/W and S/W) arose in 1985, only 6 years before this paper. Before that, floating point representations were often proprietary (e.g. Cray or IBM). Back then, most microprocessors did not have floating point hardware, so most non-scientific apps avoided floating point operations and booleans using them. Few coders had much experience with floating point unless they were engineers or scientists. This paper broke new-ish ground in introducing 'the rest of us' to the nonlinear representation space of digital scientific numbers and subtleties of roundoff & inaccuracy inherent in math-intensive libraries and scientific math functions. It still usefully covers more ground in that space than most of today's BSCS degree programs.