The Subway Doesn't Have to Be Hot and Gross

The major IRT lines save for the 7 operate mixed fleets of NTT and R62/A equipment, none of which share door positions IIRC. The Eastern Division still uses old rolling stock like R42s from time to time. The last time they tried getting rid of cars prematurely, the R179 order came in late and they were forced to run what cars remained had more aggressively, degrading reliability.

> You could make the test station one of the L train stations that's going to be closed for two years and install them at a super leisurely pace. (The L is basically completely isolated from the rest of the system, and already requires special rolling stock for CBTC. And it's going to be closed for two years for other repairs.)

This actually was a proposal. It was recently killed in favor of bringing elevators to Union Square, a major subway complex that is not ADA accessible. Which is really why they haven't caught on as a serious engineering proposal; the list of things that would be much better value for money is long, such as systemwide ADA accessibility, CBTC signalling, replacement of rolling stock, refurbishment of dilapidated, water-damaged stations, etc. You could easily come up with $50B worth of upgrades, extensions, etc. before platform doors became a priority. And they're not even that feasible.

> There are just a lot more doors in the station than on each train.

Got a source for this? This is intriguing because I've never heard of something like this. But this also introduces the complexity that the right doors have to open at the right time, lest a blind person walk into a gap between trains and die. Which is a lot to ask from an agency that can barely maintain something as mundane as elevators and escalators.

> IRT and BMT Eastern Division

I read this as “IRT Eastern Division and BMT Eastern Division”, and was very confused what the IRT Eastern Division was, but I think you mean all of the IRT? Indeed, the IRT (except R188) would require similar door locations already because of the moving platforms at some stations.

> I've never even heard it seriously proposed.

There were apparently budgeted funds in the 2014 Capital Plan [1], but it was just designs.

[1] http://web.mta.info/capitaldashboard/allframenew_head.html?P...

I have seen platform doors arranged like what you described at an inter-city railway station. Unfortunately this is unlikely to work for hot and gross subway systems like the NYC subway, where platforms tend to be narrow and overcrowded.

That would neither really restrict airflow to the point where AC would be feasible, or prevent passengers from falling onto the tracks. Not sure what that would solve as a problem.

The separate fleets are assigned to individual lines (which are better thought of as services). But there are 36 of them, pretty much all of which interweave with each other. The R, for example, at some point shares track with the N, the W, and the M, which in turn share track with the Q, the D, and the F, which in turn share track with the E, the B, the G, and so on and so forth. And this is before we consider the fact that in New York, trains are often rerouted due to service changes onto tracks they don't normally serve, and sometimes are changed after years of change in ridership patterns; installing platform doors incompatible with certain trains would reduce operational flexibility.

Untangling the lines is not possible, partially because the system was designed to provide as many one seat rides as possible, and partially because the current subway is a patchwork mashup of three separate, competing subway systems; if you were to separate the lines, some trips would just not be possible via transfers, and most transfer complexes would not be able to handle the amount of passengers changing trains.

To answer how many cars are not compatible, that would be 2025 subway cars out of a total of 6148 (at least according to Wikipedia.) Of these, 625 were built in the late '80s and are not due for retirement til the 2030s. To put that in perspective, the late '80s fleet is bigger than the entire Boston T rapid transit fleet (Green and Silver Lines not withstanding)

People will end up in this no-mans-land with the train gone, and the doors closed?

In many places there isn't "a few feet" available from the platform.

Yes, and I mentioned that (and the same article) in my post :). But even if you ignore NYC being more expensive, the more recent estimated per-station prices in Paris and Montreal would come to around US$1.5–3 billion for all of NYC’s stations. If NYC remains around 6x the price in Paris, then that’s around US$9 billion for platform doors in NYC.

Right. I traveled once on a long-distance bus line that acted like a zero-gravity plane for a split second, as it went over a bump in the road. Just wondering if the experience would be comparable :)

Maybe this could be solved by having platform doors that open in all the required positions?

You could either add extra platform doors or make platform doors that are wider than the train doors and open part way to match the train.

Unfortunately, if cars have different lengths, it increases permutations a lot. But you could handle even that by building a wall made entirely of doors (staggered on two door tracks), so the gap between any two doors can open up to any desired width.

It is still doable, you can start by splitting the running stock by similarly aligned doors run them on few line and start with the few lines which are now already aligned.

Realistically by the time you are even close to half the 470 stations it will be late 2020's anyway. Then you can aim for full coverage as the stock is deprecated.

It is more likely the $1-2 million per station cost is not high priority as say CBTC which will improve revenue as it increases capacity.