John Lees' "Reaction from the Trenches" (SIGDOC Vol. 7, No. 2) has raised some interesting points on the relationship between documentation and a concept of "good" programs. I have worked on some large projects for both systems programming and applications programming jobs over the years. In recent years I have taught a few programming courses for beginners and more advanced programmers. This led me to explore firstly, the way that programs are developed - both good and bad, and then to try to discover how good people should be introduced and guided toward a notion of "good" programming practice. I shall describe how these concepts evolved and led to a view which requires most documentation to be produced before any code is written. Indeed, it becomes part of the preparation and planning for coding and then testing. Thus, documentation is no longer seen as something separate from the software and as an evil extra that prevents programmers from getting on with the next interesting and important program. Rather it is seen as an essential and integral part of a well organized approach to producing software products. Finally, I believe that it is a moot point as to whether it is worse to have software which is poorly documented or software that does not work.
/lp/association-for-computing-machinery/literature-from-the-software-tarpit-or-the-how-where-and-why-of-IdfdX9QgNy