no silver bullet
expressionprogrammingops culture
Fred Brooks's claim that no single technique will yield an order-of-magnitude gain in software productivity, because complexity is essential, not accidental.
Brooks's 1986 essay split software difficulty into accidental complexity, which tools can remove, and essential complexity, the irreducible intricacy of the problem itself, and predicted no tenfold silver bullet within a decade. Four decades of languages, methodologies, and now AI assistants have each been announced as the exception. The essay endures as the standard rebuttal to every 'this changes everything' launch, and the standing challenge each one must answer.
Fred Brooks argued in 1986 that no single development in the following decade would produce an order-of-magnitude improvement in software productivity, and the argument rests on a distinction that is still the most useful one in the field: essential complexity versus accidental complexity.
Accidental complexity is the difficulty introduced by our tools: memory management, build systems, deployment, language ceremony. It can be attacked, and it has been, substantially. Essential complexity is the difficulty of the problem itself, of understanding what is actually required, of the interactions in a system that models something genuinely complicated. No tool removes it, because it is not in the tool.
The prediction has aged remarkably well, and each generation tests it again. Object orientation, visual programming, CASE tools, agile methods, cloud platforms and now code generation have each been proposed as the exception. Each delivered real improvements to accidental complexity and none removed the need to understand the problem. Whether the current generation is different is a live argument, and Brooks's framing is still the right one to have it in.