6 ms·

This is insane. These device UIs should be standardized and any manufacturer making a significant UI change to their product should not be permitted to sell it

This is insane. These device UIs should be standardized and any manufacturer making a significant UI change to their product should not be permitted to sell it under the same name and number. The FDA should be looking at this before approving such devices for use. Don't aircraft have more or less standardized controls and displays?

giantg2 4y ago
"Don't aircraft have more or less standardized controls and displays?" Commercial pilots (and military) are required to train and qualify on a particular aircraft before being allowed to fly it. Controls can differ, but it's generally the difference in flight characteristics/handling that is the difference. Although we can see that the different computer related controls and lack of training were a factor in the Max crashes.
- ncmncm 4y ago
  The main cause was completely insane specifications. With a reasonable design the plane would not have become uncontrollable under any possible sequence of events, training or no, familiarity or no.
  - giantg2 4y ago
    True, but we live in an age where we we trust our lives with computers so that we can do things we couldn't do otherwise. Flight control computers are necessary for most new commercial and military aircraft because they're fly by wire or can be uncontrollable by nature (Max, V22, etc). We're at the point that the FAA and people in general don't care about an inherent instability if the computer can fly it.
    - ncmncm 4y ago
      The computer was not flying it. The computer was actively fighting the pilot, who was desperately trying to keep it flying, while the computer was forcing it to crash instead. To be clear: with the computer entirely powered off, the plane would have flown just fine. Not exactly like other 737s, but more like them than anything else.
      - giantg2 4y ago
        "The computer was not flying it. ... the computer was forcing it to crash instead." Sounds to me like the computer was in fact in control of the plane if it causes it to crash. The "computer" is the MCAS system that sits between the pilot and the actual control surfaces. This is a fly by wire aircraft. Yes they could have turned the MCAS off and had it fly ok... if they even knew it existed. Although they weren't trained on the different flight characteristics they would experience with it off either. "with the computer entirely powered off, the plane would have flown just fine." Just to be clear, if you had a total computer failure or power down, a fly by wire aircraft could not function. I assume you are only talking about the MCAS.
        ncmncm 4y ago
        No, 737MAX is not a fly-by-wire aircraft. The controls are physically connected to control surfaces via actual physical cables. But, the computer can yank on those cables, and it can also run the elevator trim up and down, and does. So, no, I am not just talking about "the MCAS". There is no such thing. That is a software feature that is part of the autopilot system, not any sort of hardware. They could have pulled the circuit breaker on the whole flight computer, and flown without it.
- dTal 4y ago
  Ah yes, the training on what to do if the computer suddenly decides to crash your airplane out of the blue. I feel this framing mis-allocates blame away from Boeing and towards airlines. The MAX was specifically designed to not require a new type rating (they have now introduced one). And no amount of training is a reasonable substitute for the autopilot with a habit of nosediving into the ground because you were too cheap to build fault-checking into your AOA sensor subsystem (an oversight that should have made the MAX fail certification).
  - giantg2 4y ago
    "And no amount of training is a reasonable substitute for the autopilot with a habit of nosediving" If they would know that the system even existed then maybe they would have known how to disengage it. Yes, the design was flawed. Even then, you train for equipment failures because everything can break. It's hard to do that if your not even told that the MCAS exists. The training aspect still squarely places blame on Boeing and the FAA for crearing/approving a design that didn't mandate additional training.
    - dTal 4y ago
      My memory is fuzzy here, but wasn't it the case that even with training, you only had a window of a few seconds to disengage the MCAS before the aircraft became unrecoverable? If you waited too long, it cranked the jackscrew to its extreme position, and there was no way to un-crank it with the MCAS disabled (i.e. MCAS had exclusive control over it). That seems like an unacceptably fragile failure mode to me.
  - otikik 4y ago
    > Ah yes, the training on what to do if the computer suddenly decides to crash your airplane out of the blue. My expectations might be too high, but yes I expect their training to include something along those lines. Perhaps not "computer is trying to kill you" specifically, but "computer is off or malfunctioning". And I would expect the airplane to have a secondary/backup computer at minimum, these systems are supposed to be heavily redundant after all).
pjc50 4y ago
Good way to make medical tech even less feasible and affordable.
- jcims 4y ago
  In other words, there’s no free lunch. Regulations can be incredibly expensive to support and almost always stifle innovation and increase time to bring a product to market. Sometimes it’s still worth it. The aviation industry has had to deal with this forever because its hard to hide a plane crash. But in the end we have a very reliable mode of transportation despite its inherent risks. Accidentally overdosing an elderly or critical patient? Hell the nurse that did it might not even know. I’m sure there are ways to classify products to help with this. Anything where operator error creates a life safety issue, like IV pumps, could have some pretty rigorous controls over UX modifications.
  - concordDance 4y ago
    Note that the more rigorous the regs the more likely that a particular treatment or tool will not be affordable and the more likely someone will die because there just isn't enough X to go around. There isn't a single definite answer, it depends on how expensive the regs are and how much they help avoid damage.
seren 4y ago
As far as I am aware, for every significant UI change we do at the medical device company I am working for, we bring a panel of final users (doctors and techs) to our facility and we have to do some validation tests, and finally show these validation results to the FDA before releasing the product.
- meragrin_ 4y ago
  > we bring a panel of final users (doctors and techs) to our facility Are they the same group of people each time?
  - seren 4y ago
    I am not really involved with these tests: I believe they are not the same people every time, because doctors tend to be quite busy, and don't have much free time available. They are coming at least from different continents, but obviously to answer the sibling's comment they are selected by the company, from a pool of advanced users or early adopters, so not necessarily a good reflection of the whole population of final users. And this is an issue in itself because the product is influenced by expert users, but at least, you cannot release a totally horrible or misleading UI that no user has ever seen before. The other side effect, intended or not, is that you cannot change the UI at every iteration of the product because these tests tends to be costly, as a result, the evolution of the design is quite conservative, which might also explain in part why medical devices UI tends to look "outdated".
- couchand 4y ago
  Is your panel of end users self-selected, by chance?
- spaetzleesser 4y ago
  "As far as I am aware, for every significant UI change we do at the medical device company I am working for, we bring a panel of final users (doctors and techs) to our facility and we have to do some validation tests, and finally show these validation results to the FDA before releasing the product." I also work for a device company and a lot of these panel members aren't really competent at giving good feedback because they lack the technical knowledge. And the FDA mostly just checks for completeness of paperwork. They look at the actual product only once there is a problem in the field.
danachow 4y ago
> Don't aircraft have more or less standardized controls and displays? No, not really. There are similarities in avionics, bare minimum instrumentation and historical conventions but there are type ratings for commercial pilots - they train on each one specifically. Some aircraft are similar enough that they have “common type ratings”. But an Airbus and a Boeing look very different.
- jaywalk 4y ago
  Boeing 737NG cockpit: https://cdn.wallpapersafari.com/16/15/2AyKUB.jpg https://cdn.wallpapersafari.com/16/15/2AyKUB.jpg Airbus A320 cockpit: https://airbus-h.assetsadobe2.com/is/image/content/dam/products-and-solutions/interior/a320-cockpit.jpg?wid=1280&qlt=85,0 https://airbus-h.assetsadobe2.com/is/image/content/dam/produ... They aren't very different. For normal flying, the biggest difference is that Boeings have yokes and Airbus uses sidesticks. The important stuff is all roughly the same.
  - danachow 4y ago
    The yoke/sidestick is really a superficial difference. Though I will also point out that anybody that is good at the 10 differences puzzles will call bullshit on your claim they aren’t very different - in appearance at least - just look at the primary flight display. More relevant, look at the autopilot stack - they not only appear different they have some deeper fundamental differences in operation.
    - jaywalk 4y ago
      Oh come on. The PFDs are damn near identical. The biggest difference is that the Airbus attitude indicator has rounded corners while Boeing's is squared. That's an entirely superficial difference that changes nothing about how it works or is interpreted.
      - danachow 4y ago
        Yes, as stated, just like with the side stick/yoke it’s a superficial difference - but since you were making your case with pictures - whether they look similar or not really has to do with how detail oriented you are. Though you ignore the relevant point made about the autopilots. The entire workflow of the flight computers differs. And ask any pilot type certified in both, they typically have preferences. I’m not arguing these need to be made more standard but there’s a continuum where at one extreme the standardization is so rigid that muscle memory transfers from type to type. That’s not the case here at all.

giantg2 4y ago

"Don't aircraft have more or less standardized controls and displays?" Commercial pilots (and military) are required to train and qualify on a particular aircraft before being allowed to fly it. Controls can differ, but it's generally the difference in flight characteristics/handling that is the difference. Although we can see that the different computer related controls and lack of training were a factor in the Max crashes.

ncmncm 4y ago
The main cause was completely insane specifications. With a reasonable design the plane would not have become uncontrollable under any possible sequence of events, training or no, familiarity or no.
- giantg2 4y ago
  True, but we live in an age where we we trust our lives with computers so that we can do things we couldn't do otherwise. Flight control computers are necessary for most new commercial and military aircraft because they're fly by wire or can be uncontrollable by nature (Max, V22, etc). We're at the point that the FAA and people in general don't care about an inherent instability if the computer can fly it.
  - ncmncm 4y ago
    The computer was not flying it. The computer was actively fighting the pilot, who was desperately trying to keep it flying, while the computer was forcing it to crash instead. To be clear: with the computer entirely powered off, the plane would have flown just fine. Not exactly like other 737s, but more like them than anything else.
    - giantg2 4y ago
      "The computer was not flying it. ... the computer was forcing it to crash instead." Sounds to me like the computer was in fact in control of the plane if it causes it to crash. The "computer" is the MCAS system that sits between the pilot and the actual control surfaces. This is a fly by wire aircraft. Yes they could have turned the MCAS off and had it fly ok... if they even knew it existed. Although they weren't trained on the different flight characteristics they would experience with it off either. "with the computer entirely powered off, the plane would have flown just fine." Just to be clear, if you had a total computer failure or power down, a fly by wire aircraft could not function. I assume you are only talking about the MCAS.
      - ncmncm 4y ago
        No, 737MAX is not a fly-by-wire aircraft. The controls are physically connected to control surfaces via actual physical cables. But, the computer can yank on those cables, and it can also run the elevator trim up and down, and does. So, no, I am not just talking about "the MCAS". There is no such thing. That is a software feature that is part of the autopilot system, not any sort of hardware. They could have pulled the circuit breaker on the whole flight computer, and flown without it.
dTal 4y ago
Ah yes, the training on what to do if the computer suddenly decides to crash your airplane out of the blue. I feel this framing mis-allocates blame away from Boeing and towards airlines. The MAX was specifically designed to not require a new type rating (they have now introduced one). And no amount of training is a reasonable substitute for the autopilot with a habit of nosediving into the ground because you were too cheap to build fault-checking into your AOA sensor subsystem (an oversight that should have made the MAX fail certification).
- giantg2 4y ago
  "And no amount of training is a reasonable substitute for the autopilot with a habit of nosediving" If they would know that the system even existed then maybe they would have known how to disengage it. Yes, the design was flawed. Even then, you train for equipment failures because everything can break. It's hard to do that if your not even told that the MCAS exists. The training aspect still squarely places blame on Boeing and the FAA for crearing/approving a design that didn't mandate additional training.
  - dTal 4y ago
    My memory is fuzzy here, but wasn't it the case that even with training, you only had a window of a few seconds to disengage the MCAS before the aircraft became unrecoverable? If you waited too long, it cranked the jackscrew to its extreme position, and there was no way to un-crank it with the MCAS disabled (i.e. MCAS had exclusive control over it). That seems like an unacceptably fragile failure mode to me.
- otikik 4y ago
  > Ah yes, the training on what to do if the computer suddenly decides to crash your airplane out of the blue. My expectations might be too high, but yes I expect their training to include something along those lines. Perhaps not "computer is trying to kill you" specifically, but "computer is off or malfunctioning". And I would expect the airplane to have a secondary/backup computer at minimum, these systems are supposed to be heavily redundant after all).

pjc50 4y ago

Good way to make medical tech even less feasible and affordable.

jcims 4y ago
In other words, there’s no free lunch. Regulations can be incredibly expensive to support and almost always stifle innovation and increase time to bring a product to market. Sometimes it’s still worth it. The aviation industry has had to deal with this forever because its hard to hide a plane crash. But in the end we have a very reliable mode of transportation despite its inherent risks. Accidentally overdosing an elderly or critical patient? Hell the nurse that did it might not even know. I’m sure there are ways to classify products to help with this. Anything where operator error creates a life safety issue, like IV pumps, could have some pretty rigorous controls over UX modifications.
- concordDance 4y ago
  Note that the more rigorous the regs the more likely that a particular treatment or tool will not be affordable and the more likely someone will die because there just isn't enough X to go around. There isn't a single definite answer, it depends on how expensive the regs are and how much they help avoid damage.

seren 4y ago

As far as I am aware, for every significant UI change we do at the medical device company I am working for, we bring a panel of final users (doctors and techs) to our facility and we have to do some validation tests, and finally show these validation results to the FDA before releasing the product.

meragrin_ 4y ago
> we bring a panel of final users (doctors and techs) to our facility Are they the same group of people each time?
- seren 4y ago
  I am not really involved with these tests: I believe they are not the same people every time, because doctors tend to be quite busy, and don't have much free time available. They are coming at least from different continents, but obviously to answer the sibling's comment they are selected by the company, from a pool of advanced users or early adopters, so not necessarily a good reflection of the whole population of final users. And this is an issue in itself because the product is influenced by expert users, but at least, you cannot release a totally horrible or misleading UI that no user has ever seen before. The other side effect, intended or not, is that you cannot change the UI at every iteration of the product because these tests tends to be costly, as a result, the evolution of the design is quite conservative, which might also explain in part why medical devices UI tends to look "outdated".
couchand 4y ago
Is your panel of end users self-selected, by chance?
spaetzleesser 4y ago
"As far as I am aware, for every significant UI change we do at the medical device company I am working for, we bring a panel of final users (doctors and techs) to our facility and we have to do some validation tests, and finally show these validation results to the FDA before releasing the product." I also work for a device company and a lot of these panel members aren't really competent at giving good feedback because they lack the technical knowledge. And the FDA mostly just checks for completeness of paperwork. They look at the actual product only once there is a problem in the field.

danachow 4y ago

> Don't aircraft have more or less standardized controls and displays? No, not really. There are similarities in avionics, bare minimum instrumentation and historical conventions but there are type ratings for commercial pilots - they train on each one specifically. Some aircraft are similar enough that they have “common type ratings”. But an Airbus and a Boeing look very different.

jaywalk 4y ago
Boeing 737NG cockpit: https://cdn.wallpapersafari.com/16/15/2AyKUB.jpg https://cdn.wallpapersafari.com/16/15/2AyKUB.jpg Airbus A320 cockpit: https://airbus-h.assetsadobe2.com/is/image/content/dam/products-and-solutions/interior/a320-cockpit.jpg?wid=1280&qlt=85,0 https://airbus-h.assetsadobe2.com/is/image/content/dam/produ... They aren't very different. For normal flying, the biggest difference is that Boeings have yokes and Airbus uses sidesticks. The important stuff is all roughly the same.
- danachow 4y ago
  The yoke/sidestick is really a superficial difference. Though I will also point out that anybody that is good at the 10 differences puzzles will call bullshit on your claim they aren’t very different - in appearance at least - just look at the primary flight display. More relevant, look at the autopilot stack - they not only appear different they have some deeper fundamental differences in operation.
  - jaywalk 4y ago
    Oh come on. The PFDs are damn near identical. The biggest difference is that the Airbus attitude indicator has rounded corners while Boeing's is squared. That's an entirely superficial difference that changes nothing about how it works or is interpreted.
    - danachow 4y ago
      Yes, as stated, just like with the side stick/yoke it’s a superficial difference - but since you were making your case with pictures - whether they look similar or not really has to do with how detail oriented you are. Though you ignore the relevant point made about the autopilots. The entire workflow of the flight computers differs. And ask any pilot type certified in both, they typically have preferences. I’m not arguing these need to be made more standard but there’s a continuum where at one extreme the standardization is so rigid that muscle memory transfers from type to type. That’s not the case here at all.