> "The primary goal, to personally obtain first-hand experience, and to reach full understanding of every detail, inherently determined our manpower: two part-time programmers. We tentatively set our time-limit for completion to three years. As it later turned out, this had been a good estimate; programming was begun in early 1986, and a first version of the system was released in the fall of 1988."

I think of Wirth and Gutknecht as demonstrating that one need not be as far out as Terry A. Davis to do small-team end-to-end work. (compare Carver Mead's tall thin person)