3 ms·

This is a pretty cavalier way to look at engineering standards. The majority of regulations and standards in consumer/public facing engineering works are writt

by abduhl 6y ago

This is a pretty cavalier way to look at engineering standards. The majority of regulations and standards in consumer/public facing engineering works are written in blood. Saying that we "tried a bunch of different ones and found one that worked best" does a disservice to the people that died for us to learn our lessons. Retreading the cowboy attitude that led to these regulations is not something that should be encouraged.

imgabe 6y ago
Suppose cars didn't have headlights, how would you go about finding the right height to mount them at? You can do calculations, test in a lab, etc, but eventually you have to put them on a car and ship them. In the meantime, how many deaths would you allow due to cars not being able to drive at night? There is no scenario in which nobody dies. Doing something, even if it's not perfect, is often better than doing nothing.
- abduhl 6y ago
  Suppose a 737 had engines that were old and not as efficient as newer, larger engines. How would you go about finding the right changes to the airframe to accommodate these new engines? You can do calculations, tests in a lab, etc, but eventually you have to put them on a plane and ship them. Oops. But then again, doing something is often better than nothing. Even if it’s not perfect.
  - imgabe 6y ago
    Cover-up of a known flaw is not the same thing. You do eventually have to ship something, no test is ever going to uncover 100% of all possible problems. I guess you'd prefer to just never release anything? Then, certainly humans will live forever.
    - abduhl 6y ago
      The point is what the GP said: “It's a way better choice to be in a future where we are arguing about the extent and content of the regulation rather than the presence of it in the first place. A deregulated world is scary to me.” If you believe a deregulated world is better then you should be pleased with the progress made by engineers and managers like those at Boeing involved with the 737 who disrupted and circumvented the regulatory state for airplanes.
      - imgabe 6y ago
        There are more choices than "no regulations" and "every regulation". Regulations created to address proven problems are one thing. Creating a new regulation because somebody dreamed up a theoretical scenario is a different thing. Progress involves risk. Cowering and doing nothing because someone, somewhere might possibly be harmed is as bad as having no regulation, if not worse. We wouldn't have airplanes at all if someone were running after the Wright brothers telling them it's not safe. How does "Maybe we don't need 800 bazillion regulations that make it impossible to do anything" get translated to "Oh, so you want a wild west cowboy world where everyone does whatever they want and there are no regulations at all?" Try and have at least a little bit of nuance. Neither I nor TFA are arguing to abolish all regulations. Just recognize that regulation has a cost. It's not an unalloyed good. When speed matters, like getting a vaccine to end a global pandemic, that cost is measured in lives just as surely as the cost of a lack of regulation is.
winston_smith 6y ago
The key is correctly assessing what part of the life cycle you're in. If you're building a bridge for millions of cars to drive on, that's something we've done many times and which needs to be done correctly and carefully, so you should definitely pay attention to all the carefully collected wisdom embodied in standards. If you're inventing clever new algorithms, then you should experiment as efficiently as possible, because 99% of what you try will be thrown away. Carefully planned waterfall style development will actively hurt you at this stage, because you have no idea where you're going yet. If you find something promising and want to scale it, then you slow down a little and check your work. This continues as you scale, and by the time your billion dollar business depends on your clever algorithm, the update process for your production version should look a lot more like the bridge building process.

by abduhl 6y ago

imgabe 6y ago

Suppose cars didn't have headlights, how would you go about finding the right height to mount them at? You can do calculations, test in a lab, etc, but eventually you have to put them on a car and ship them. In the meantime, how many deaths would you allow due to cars not being able to drive at night? There is no scenario in which nobody dies. Doing something, even if it's not perfect, is often better than doing nothing.

abduhl 6y ago
Suppose a 737 had engines that were old and not as efficient as newer, larger engines. How would you go about finding the right changes to the airframe to accommodate these new engines? You can do calculations, tests in a lab, etc, but eventually you have to put them on a plane and ship them. Oops. But then again, doing something is often better than nothing. Even if it’s not perfect.
- imgabe 6y ago
  Cover-up of a known flaw is not the same thing. You do eventually have to ship something, no test is ever going to uncover 100% of all possible problems. I guess you'd prefer to just never release anything? Then, certainly humans will live forever.
  - abduhl 6y ago
    The point is what the GP said: “It's a way better choice to be in a future where we are arguing about the extent and content of the regulation rather than the presence of it in the first place. A deregulated world is scary to me.” If you believe a deregulated world is better then you should be pleased with the progress made by engineers and managers like those at Boeing involved with the 737 who disrupted and circumvented the regulatory state for airplanes.
    - imgabe 6y ago
      There are more choices than "no regulations" and "every regulation". Regulations created to address proven problems are one thing. Creating a new regulation because somebody dreamed up a theoretical scenario is a different thing. Progress involves risk. Cowering and doing nothing because someone, somewhere might possibly be harmed is as bad as having no regulation, if not worse. We wouldn't have airplanes at all if someone were running after the Wright brothers telling them it's not safe. How does "Maybe we don't need 800 bazillion regulations that make it impossible to do anything" get translated to "Oh, so you want a wild west cowboy world where everyone does whatever they want and there are no regulations at all?" Try and have at least a little bit of nuance. Neither I nor TFA are arguing to abolish all regulations. Just recognize that regulation has a cost. It's not an unalloyed good. When speed matters, like getting a vaccine to end a global pandemic, that cost is measured in lives just as surely as the cost of a lack of regulation is.

winston_smith 6y ago

The key is correctly assessing what part of the life cycle you're in. If you're building a bridge for millions of cars to drive on, that's something we've done many times and which needs to be done correctly and carefully, so you should definitely pay attention to all the carefully collected wisdom embodied in standards. If you're inventing clever new algorithms, then you should experiment as efficiently as possible, because 99% of what you try will be thrown away. Carefully planned waterfall style development will actively hurt you at this stage, because you have no idea where you're going yet. If you find something promising and want to scale it, then you slow down a little and check your work. This continues as you scale, and by the time your billion dollar business depends on your clever algorithm, the update process for your production version should look a lot more like the bridge building process.