In regard to "Young promises acquisition changes," the idea of using prototypes to test designs and manufacturing processes has been around for a long time.

Somehow, perhaps in a rush to quick implementation or penny-wise and pound-foolish attempts to save money, prototyping went out of practice for large Defense Department systems. It's good to see it coming back. "Measure twice, cut once" is a philosophy that shouldn't be ignored, especially where important sums of money are at stake or important decisions are being made.

Thirty years ago in a paper titled "Model Centered Management" I tried to make the case for using models (analytical models, simulation models, prototypes) to give managers an "artificial," third-party testbed to evaluate computer system designs and implementation technologies. There were too many large-system implementation failures, and I believed then, and still do, that there is no substitute for experience with a working system -- even if it is just modeled -- to test policies and procedures, technologies and processes, algorithms and heuristics, that although they may look good on paper, have not been tried and tested.

The DOD prototyping policy is long overdue. I'd like to see it adapted to nonweapons systems environments. As for the extra cost and extra time involved, why do you think carpenters invested the "measure twice, cut once" rule?

Anonymous