6 ms·

Do small things first -- make the countdown clocks work. It will go a long way towards making people believe MTA digital projects. Make sure that when the elec

Do small things first -- make the countdown clocks work. It will go a long way towards making people believe MTA digital projects. Make sure that when the electronic sign says that the train is "2 minutes away" it actually shows up in 2 minutes. If a sign says train is "6 minutes away" there should not be a train entering a platform 2 minutes later.

VWWHFSfQ 7y ago
maybe it's just my train, but the countdown clocks have always been pretty accurate for me.
- Tokkemon 7y ago
  Yes they are on the whole. But that one time there was some unexpected delay, and it ruins any goodwill the clock built up in the eyes of the rider. It's quite frustrating to see.
  - jrockway 7y ago
    I have seen them wrong, it didn't ruin the goodwill. What I take away from the clocks are minor delays that don't get mentioned in the official notifications. I often see arrival times like "1 3 15 21" (literally a sample taken right now for my home station) and know something is wrong, but there is no communication from the MTA about it. I then choose an alternate route. What the MTA needs to fix is when the schedule says that trains run every 3 minutes, the countdown clock needs to look like "1 4 7 10", not "1 4 33 36".
  - notyourday 7y ago
    The purpose of the countdown clock is to provide information about that "unexpected delay". The clock is irrelevant if the trains are arriving every 3-4 minutes. Think of this being an autoscaling group. The argument that people make sounds like "Well, autoscaling group works fine when one instance can handle all the traffic. We are happy with it." It is not an achievement to have an autoscaling group work when it has one instance in it.
ixtli 7y ago
If you think “make the countdown clocks work” is a small thing you haven’t researched this problem thoroughly. https://www.theatlantic.com/technology/archive/2015/11/why-dont-we-know-where-all-the-trains-are/415152/ https://www.theatlantic.com/technology/archive/2015/11/why-d...
- notyourday 7y ago
  Yes, it is a very small thing. Start with automated lines ( the L ) where you know the location of every train. You know the speed of every train at every point. Now it is a standard algebra problem that 15 year olds were solving 20 years ago in algebra/comp sci exams, using pen and paper - it was pre-calc so they invented how they could divide the track into equal section and compute the time it takes the train it should take the train to get to the station by approximating the unknown time based off the time that it took the train to pass the sections behind it. On non-automated lines start by adding god damn train trackers to the individual trains and start splitting the blind sections of the track by adding cameras and sensors. They don't need to make the lines automatic. They just need to add visibility. None of this is difficult. MTA and "infrastructure is hard" people are full of excuses. They are the same people who were saying that the L should be fully shutdown for 18 months when now we have MTA announcement that even with the partial shutdown they are way ahead of schedule with the tunnel work so they can start doing the station work. There should never be the case when it says a train is 8 minutes away and a train behind it is 9 minutes away when a train gets to the station in 2 minutes and the train behind it gets to the station in 5 minutes.
  - tialaramex 7y ago
    What you've done there is you've found a very complicated way to solve entirely the wrong problem. Bravo, I guess.
    - notyourday 7y ago
      "I hate the way our application is written. We need to be able to instrument it and also rewrite all that icky old logic. Basically we should rewrite the entire thing. It will take us about 3 years"

VWWHFSfQ 7y ago

maybe it's just my train, but the countdown clocks have always been pretty accurate for me.

Tokkemon 7y ago
Yes they are on the whole. But that one time there was some unexpected delay, and it ruins any goodwill the clock built up in the eyes of the rider. It's quite frustrating to see.
- jrockway 7y ago
  I have seen them wrong, it didn't ruin the goodwill. What I take away from the clocks are minor delays that don't get mentioned in the official notifications. I often see arrival times like "1 3 15 21" (literally a sample taken right now for my home station) and know something is wrong, but there is no communication from the MTA about it. I then choose an alternate route. What the MTA needs to fix is when the schedule says that trains run every 3 minutes, the countdown clock needs to look like "1 4 7 10", not "1 4 33 36".
- notyourday 7y ago
  The purpose of the countdown clock is to provide information about that "unexpected delay". The clock is irrelevant if the trains are arriving every 3-4 minutes. Think of this being an autoscaling group. The argument that people make sounds like "Well, autoscaling group works fine when one instance can handle all the traffic. We are happy with it." It is not an achievement to have an autoscaling group work when it has one instance in it.

ixtli 7y ago

If you think “make the countdown clocks work” is a small thing you haven’t researched this problem thoroughly. https://www.theatlantic.com/technology/archive/2015/11/why-dont-we-know-where-all-the-trains-are/415152/ https://www.theatlantic.com/technology/archive/2015/11/why-d...

notyourday 7y ago
Yes, it is a very small thing. Start with automated lines ( the L ) where you know the location of every train. You know the speed of every train at every point. Now it is a standard algebra problem that 15 year olds were solving 20 years ago in algebra/comp sci exams, using pen and paper - it was pre-calc so they invented how they could divide the track into equal section and compute the time it takes the train it should take the train to get to the station by approximating the unknown time based off the time that it took the train to pass the sections behind it. On non-automated lines start by adding god damn train trackers to the individual trains and start splitting the blind sections of the track by adding cameras and sensors. They don't need to make the lines automatic. They just need to add visibility. None of this is difficult. MTA and "infrastructure is hard" people are full of excuses. They are the same people who were saying that the L should be fully shutdown for 18 months when now we have MTA announcement that even with the partial shutdown they are way ahead of schedule with the tunnel work so they can start doing the station work. There should never be the case when it says a train is 8 minutes away and a train behind it is 9 minutes away when a train gets to the station in 2 minutes and the train behind it gets to the station in 5 minutes.
- tialaramex 7y ago
  What you've done there is you've found a very complicated way to solve entirely the wrong problem. Bravo, I guess.
  - notyourday 7y ago
    "I hate the way our application is written. We need to be able to instrument it and also rewrite all that icky old logic. Basically we should rewrite the entire thing. It will take us about 3 years"