George Dantzig had already established his reputation as one of the most important applied mathematicians of our time (think linear programming) so he can be excused for writing a dumb book. Compact City: A plan for a liveable urban environment (1973) is one of those instances where smart person overreaches and embarrasses himself.
Just a little urban economics would have been sufficient to reveal to the great man that there are agglomeration-congestion trade-offs -- and they are peculiar to each setting. In downtown NYC, the trade-off is met via one urban form and in Silicon Valley it is met via another. One size does not fit all.
USC's Bumsoo Lee reports that in 2000, drive-alone commuters to Manhattan's CBD averaged 56 minutes each way; at the other extreme (among the largest U.S. metros) were those driving to dispersed places of employment in Phoenix (25 minutes each way).
Which is better? What are we optimizing? The point is that (fleeting) equilibria work out differently in light of history and available infrastrucutre.
Why bring all this up? Because compact settlement is still the implicit or explicit prescription in hundreds of today's New Urban plans and projects.
In that sense, Dantzig was not simply wrong but wrong and ahead of his time.