Commentary on Donald MacKenzie, “A View from the Sonnenbichl”

Abstract

The 1968 Garmisch Conference not only coined the term ‘software engineering,’ thus making it clear that the development of software is an engineering task, but it also addressed, for the first time, a profound problem of major concern—the difficulty of achieving dependable software. I can understand well that a historian and sociologist like Donald MacKenzie takes this as a starting point for considering the historical and sociological aspects of software and system dependability. But I believe that this involves the danger of taking too narrow a view, neglecting the commercial aspects of development and maintenance. These factors both inhibit and prohibit software dependability, perhaps to a much greater degree than the skills of the software engineer do. They also force us to distinguish between various types of software and to be more precise about the phenomenon of dependability and ways of achieving it.