But they aren't... the Heathrow Express underperforms ridership expectations. Most likely, the CDG Express will underperform as well, because to the 99% or so of CDG travelers who aren't going to the immediate Gare de l'Est area, the RER B will both remain faster and save them a transfer. BART to SFO not only missed projections by a factor of about 4, but also made public transit on the Peninsula worse: first, the expense of building it led SamTrans to gut the county bus service, and second, the BART extension branches between Millbrae (with a connection to Caltrain) and SFO, so suburban travelers from points south have to transfer Caltrain-BART-BART-Airtrain, none of which transfers is timed.

The station art would be horrendous, judging by their poetry.

Some large systems come close to this fleet uniformity, by making huge orders. This tends to result in lower per-unit costs. New York's most recent subway order, the R160, was 1,662 cars. Its most recent commuter rail order, the M8, was small because it's for one line with different electrification, but the previous one, the M7, was 1,172 cars. Both orders were cheap by US standards. Japan has even larger orders, and (for several other reasons as well) very low rolling stock costs; JR East's E231 and E233 Series orders are both well above 2,000 cars each. The R160, E231, and E233 were all manufactured by multiple vendors, each responsible for a portion of the order, so uniformity does not mean vendor capture.

Fleet uniformity is actually a good goal to strive for, and if I were designing a rail system for scratch I'd make sure to have it to lower maintenance costs. An established system should strive for this as well, but it can take decades to retire older cars, since unlike with planes there's generally no secondary market for used subway cars as the specs are different for each city. California High-Speed Rail is thinking in such a direction, and at least one of the international HSR operators that peer-reviewed the business plan noted its advantages, while also criticizing many other aspects of the CAHSR plan.

Two tracks carrying two services is normal in many cities around the world. In New York, nearly every track pair splits into two different services somewhere, and one track pair (the 60th Street Tunnel) splits into three. In London, whose longer stop spacing and deep-suburban service has parallels with Washington, two tracks carry 2-3 services on the subsurface lines, with some of those services branching out further in the suburbs, and on the deep-level lines three lines branch as well; only in two locations do different lines interline to give local and express service. On the Paris RER and every S-Bahn system, a central trunk line carries anywhere between 2 and 6 services. In Chicago, the two-track Loop carries five different services, with flat junctions. In Tokyo some commuter lines have four tracks with local and express service, and the busiest has eight, but that's justified by immense demand, and most central track pairs are highly branched in the suburbs because of subway and commuter rail interlining.

Branching is actually very logical on a system that's used even for outer-urban service, let alone suburban service like in Washington. You're not going to get the same demand out of Dulles or New Carrollton or Bethesda as out of Union Station, Arlington, and the Capitol. This means that to avoid overcrowding the inner segment or running lots of empty trains to suburbia, you need to either short-turn some trains or branch. (None of this means it's a bad idea to build a fourth trunk line through DC, just that the idea of giving each color its own dedicated tracks is bad.)