It imagines a network of rapid transit subways many of which flow into surface rail lines, so presumably a model similar to San Francisco or Brussels or the Boston Green Line.
(He calls it a "tram-train" system, but it's actually the opposite of the pattern used by the celebrated tram-trains in Karlsruhe, because here "train" means a subway flowing through the center of the city, while in Karlsruhe the tram portion is the city and surface train lines are used to reach outer suburbs.)
Some problems to note even in a proposed utopia: Subways that flow into streetcar lines are often a poor fit, for three reasons:
- Subways are very high capacity and streetcars usually lower, if only because you physically can't run streetcars as frequently as you can run subways. In both San Francisco and Brussels, the subway service branches onto two or three streetcar branches, and that manages the problem somewhat. But alas:
- Streetcars in mixed traffic are exposed to many causes of delay, subways are not. Since a line is only as reliable as its least reliable point, the full capacity potential of the subway can never be used because cars are flowing in from the streetcar portion at unpredictable times.
- Subways are very fast and streetcars very slow. It's rare for a demand pattern intense enough to require a subway to end so abruptly that the only thing you need are streetcars. In both San Francisco and Brussels, the feeling is that if a subway has to transition into a surface light rail form, it should ideally go into exclusive right-of-way so that it's still somewhat fast, and then perhaps transition to streetcar further out as loads are lower. (The M-line in San Francisco tries to do this.) Of course, this is not precisely what happens in either San Francisco or Brussels, because like all systems they are prisoners of the long-ago design of their infrastructure, and the city has grown in so tightly around the infrastructure at it is that it would be unimaginably expensive (in political pain as much as money) to change it.
All these issues will be familiar to transit riders in San Francisco and Brussels.
Note also that ending routes in one-way loops, as this fantasist does, is not the best practice unless the loops are very small (e.g. one block wide). That's because you need a driver break and recovery point at the end of the line, and you want the vehicle to be empty at that point. If you're ending in a large loop (planners sometimes call them "balloon loops") you have the same problem as the London Circle Line: there's never a point where everyone is off so the driver can have a rest and the vehicle can get back on schedule if it's a bit late. Even driverless systems need some spare time at end of line to recover from delays.