I made a doodle. It shows the two approaches to dealing with congestion. They show how highways can approach and enter cities. The difference is increasing CORRIDOR CAPACITY and NETWORK CAPACITY.
The top is transportation planners and their dogma. The bottom is from sane people who understand that cities truly are the wealth of nations:
The difference is that transportation planners address density (and therefore congestion) they approach it as a problem of corridor capacity (it probably doesn't help that different hierarchy of roads have different jurisdictional oversight). Since there is congestion, the only solution of course is to widen the road through the city, whatever property acquisition be damned, because the road isn't operating at level of service A. What does level of service mean? The traffic flow is smooth and uninterrupted.
Although there is a flip-side to this equation. And that is the response to transportation infrastructural changes from the real estate market. The St. Louis group fighting for a highway removal referenced a great line from 1912 city plan document for St. Louis, saying the point of city planning was "the art of arranging streets and public spaces that privately owned land might be put to best use."
I don't think there is a better way to put it.
On the other hand, there is the "context-sensitive" approach, which lets the network deal with the capacity and need for movement. Where there is more congestion, the network gets more complex and more interconnected to add capacity while maximizing route choice, flexibility, and inherent intelligence built into the user. You can adapt on a case-by-case basis and a route by route basis. It also doesn't disperse real estate value away from where congestion is the highest by funneling traffic into one big, hot, stinky, dangerous corridor. There is a built-in logic, order, and intelligence to it.
There is no such intelligence to modern traffic planning, no matter how many models or formulae are referenced. The dendritic system concentrates the bad, while dispersing the good. The reticulated, network approach concentrates the good, in densely connected, walkable areas while dispersing the bad (like high speed traffic or low intensity industrial uses to the periphery).
In studies for private clients, I have shown how a network of streets can lose virtually no total capacity by converting wide one-way streets to a more interconnected system of two-way, narrower streets. Whatever capacity loss there is from lane reduction is mitigated by the fact that the sociofugal approach with bigger roads was operating at only 37% capacity at its highest traffic areas. The reduction of lanes than allowed for increased parking and reduced right-of-way, thus yielding increased developable land for more efficient buildings on what were very narrow lots (again, due to overly wide roads). Urbanism, it's all about the win-wins. Best case scenario? All of that road capacity does fill up, but it doesn't matter? Because the place is desirable, more walkable, and denser, thus allowing for increased resilience, which was what we were worried about.
Here are the by-products to the two approaches. The choice is yours (well, not really. But if you rabble-rouse enough, you might just get some say!):