But you have to remember the city sizes was not a design choice, its was a performance necessity. Over the summer they even worked to see if they could increase city sizes, and they said flat out with the current engine it is not feasible.
So there probably are some design flaws in the engine preventing it, but they claimed it would require a heavy overhaul to fix it.
We can only hope the next SimCity really takes the lessons they learned here and apply them.
And I would say it was a fatal performance necessity. They chose early on to try to simulate every citizen... and the proceed to think making a single core/single threaded game would be the best route. There had to be a point during the programming that somebody said "hey, let's see what happens when we make the biggest city possible!" They soon realized that Glassbox could only tolerate the small city sizes of the current game.
One of two events happened. First, somebody must have spoken up remarking how the engine doesn't scale well, and that went ignored. The bosses deemed the engine and city size "good enough." Or second, nobody complained from the designers since everybody bought into the idea the next game was SimTown/SimVille and not SimCity, and that consumers would be happy with the tiny city plots.
Either way, it comes down to bad decisions that severely harmed the end product. If they were going to design a game meant to simulate thousands of individuals, wouldn't the logical choice to make the game scalable and utilize the hardware resources that most modern desktops now have? Sadly they made a bad decision, and potentially fatally wounded another great gaming series.