"No silver bullet" reloaded: retrospective on "essence and accidents of software engineering"

Reviewer: Larry Bernstein

This is a report of a panel discussion reexamining the wisdom of Fred Brooks. It’s rare that the author of a seminal article sits on a panel that evaluates his work. Brooks is a good sport, and continues to be a serious contributor to the community’s understanding. Brooks scored again when he said, “The most radical possible solution for constructing is not to construct it all.” Or, as one of my students said when describing a software engineering lesson he learned on his project, “Don’t write what is already written.” Parnas makes a pithy and insightful comment about the innovation needed to reduce software complexity. He states: The belief that we can find a silver bullet is kept alive by another myth, the belief that we have made great progress in software engineering because we are managing to build increasingly complex systems. In fact, if there is progress it will [be] because we will be building increasingly simple systems. The paper is worth reading if only to learn the lessons provided at the end by Thomas. Many of the other comments repeat what we already know (and fail to do). This is a crisp review of the state of understanding; it is a quick and worthwhile read. But, I disagree with the conviction of most of the panelists that there has been little progress in improving software productivity. Consider how long it would take to write the Amazon Web site applications with the tools available in 1960. Online Computing Reviews Service