EGNILGES@pucc.UUCP (06/05/88)
In article <8806032026.AA11586@mitre.arpa>, nigam@MITRE.ARPA writes: >A recent comp.software-eng message was complaining about a code comment that >refered to a particular book, page, and algorithm for full explaination. >This started a chain of thought: what books should we expect every >software engineer to have available; preferably own. > > ... > >Numerical Analysis: > (can't think of definitive book) I'm a professional programmer rather than a mathematician, so weight my numerical analysis choices appropriately. A new (1986) book, NUMERICAL RECIPES AND FORTRAN PROGRAMMING, might fill this spot. It is by the following quartet of authors: William H. Press, Brian P. Flannery, Saul A. Teukolsky and William T. Vetterling. Press et al. tend to make value judgements without Knuth's tendency to PROVE everything. This makes NUMERICAL RECIPES a good choice for the working software engineer, who needs to trust his references but who may not have time to read and understand proofs. Don't get me wrong. Knuth is great. But Press et al. should be chosen for a numerical reference by the working software engineer, I believe. The edition of NUMERICAL RECIPES I have uses FORTRAN, but there's an appendix of numerical recipes in Pascal and a separate, simultaneous edition of NUMERICAL RECIPES IN C.