A theoretical physicist, now information science specialist at the Naval Research Laboratory in Washington, puts his finger on the ""software crisis."" According to Shore, we simply lack the means and models by which to write complex computer programs with assurance that they are correct. What's more, we have not developed tests for accuracy beyond the kind of simple enumeration of all cases (substituting for variables) that would take exhaustive periods of time. Shore begins his discussions with a short takeout on the general problem of computer anxiety--not from an ergonomics or psychological perspective, but in terms of the poorly written ""documentation"" that accompanies the average home computer (or the expensive office version). He cites the built-in ambiguities, and built-in potential for ""crashes,"" as well as the insulting bleeps and haranguing messages that flash on the video display. Then he goes on to a sage discussion of how computer languages evolved, from machine talk to assembly/compiler tongues to FORTRAN and its kin, ""Pascal"" (which he rather likes) and other new arrivals. The title is a homely example of how a family recipe can be described through a series of algorithms, each an improvement over its predecessor. Shore's almost sermonly points are that there is art and style to writing programs; that we must subject them to the same canons of intelligibility, grace, and elegance as literature or mathematical models. A timely and unhackneyed critique.