3 ms·

> Most of what was once a "trivial" design decision has become computer-controlled beyond recognition. Indeed, it's almost caused by lazy thinking. If you need

by umvi 6y ago

> Most of what was once a "trivial" design decision has become computer-controlled beyond recognition. Indeed, it's almost caused by lazy thinking. If you need conditional behavior of a widget, the default is to reach for the almighty transistor, why spend more brain cells than that. Yet... sometimes it's better to trigger your conditional behavior on some other physical phenominan (read: gas station nozzles[0]) [0] https://en.wikipedia.org/wiki/Venturi_effect https://en.wikipedia.org/wiki/Venturi_effect

Accujack 6y ago
I think it's mostly due to the change in the makeup of the average "engineering" employee in corporations today. There's a very heavy emphasis on computer hardware and software knowledge, and much less on mechanical engineering and physical design. If all you have is a hammer, everything looks like a nail. When you want to design a door handle with specific behavior and you understand software and electronics well but mechanical design less so, you design an electronically actuated door handle and do all the features in software. It's cheaper, too, honestly, because revising a software system is inherently cheaper than the design-manufacture-construct-test loop for mechanical design. But, as automotive design advances out of the technical design stasis it's been in for decades due to the dominance/barriers to entry of the US car industry, there's going to be a need for more advanced rules regarding what needs to be manual/mechanical vs. fly by wire. A similar regulatory regime already exists in aircraft design, which could be a guiding example.
jjk166 6y ago
That's the opposite of reality. More mechanical/moving parts increases risk and complexity, not reduces it. If something is misaligned or bends or breaks or even if it's just insufficiently lubricated, you're SOL. Solid state transistors are stupid reliable in comparison. Further, implementing more than two or three levels of conditional logic with mechanisms is prohibitively expensive, whereas you can have thousands of checks in a digital system practically for free. Finally, a mechanical system can only handle situations it is designed for, upgrades are difficult and sometimes impossible to implement. On the other hand digital logic is easily updated to deal with unforeseen issues. Electricity is just as much a physical phenomenon as fluid flow or mechanical stress. [Edit] I'm a mechanical engineer, no one should be any more enthusiastic about doing things mechanically than I am.
- TheCapeGreek 6y ago
  Sure, because taking out the digital system to repair/upgrade it (unless your tap is somehow wireless too) is totally different from taking out these damaged/upgradeable mechanical mechanisms.
  - jjk166 6y ago
    If you don't see a difference between updating software and hardware, I don't really know what to say.
- umvi 6y ago
  I guess the point is... say you use a microchip connected to a sensor to determine when to switch off a gas nozzle. It would work, of course, but if the sensor fails, or there's a bug in the firmware, or if power fails... gas spills everywhere. Whereas if you couple the trigger to a physical phenomenon that doesn't need electricity to work, now there's only one thing that can fail, and that's the part converting the physical phenomenon to a mechanical action. I love computers as much as anyone on this site, but I like mechanical things too. I much prefer mechanical locks to digital ones. Mechanical locks are extremely reliable and well made ones work for centuries. Whereas I wouldn't trust a modern fingerprint padlock at all - what if the sensor fails? What if the batteries run out? What if moisture gets inside and corrodes the circuit board? One of the main jobs of a lock is to open when I need it to, and I just don't trust battery powered locks to do that.
  - jjk166 6y ago
    > now there's only one thing that can fail That's the sticking point. People tend not to think of mechanical systems as systems full of many components any one of which can fail, but they are. In the case of the gas nozzle, you have the orifice for the airflow which can get blocked, you have the diaphragm which can break, you have the connection to the lever which can sever, you have the lever itself which can bind. Critically if it does fail, it fails catastrophically - gas spills everywhere. On the other hand consider a system employing a capacitive level sensor. When the sensor comes into contact with fluid, the dialectric constant changes and thus the voltage drop across the sensor increases to the point the valve will close itself. If anything goes wrong, say you lose power or a wire breaks or someone physically rips the nozzle off, the voltage still decreases below the cutoff threshold so the valve still closes - it fails safe. People like mechanical systems because they are easy to visualize, but that's not the same as being simple or safe. At the same time, most of our experience with electronics is with low cost consumer electronics where the price of failure is low and there is strong financial incentive to use cheap components with short life expectancies in configurations that aren't fail-safe, so we tend to associate electronics with being unreliable. But on a fundamental level, the force stopping two gears from passing through each other is exactly the same as the one producing a current when you apply a voltage to a piece of copper.

by umvi 6y ago

Accujack 6y ago

I think it's mostly due to the change in the makeup of the average "engineering" employee in corporations today. There's a very heavy emphasis on computer hardware and software knowledge, and much less on mechanical engineering and physical design. If all you have is a hammer, everything looks like a nail. When you want to design a door handle with specific behavior and you understand software and electronics well but mechanical design less so, you design an electronically actuated door handle and do all the features in software. It's cheaper, too, honestly, because revising a software system is inherently cheaper than the design-manufacture-construct-test loop for mechanical design. But, as automotive design advances out of the technical design stasis it's been in for decades due to the dominance/barriers to entry of the US car industry, there's going to be a need for more advanced rules regarding what needs to be manual/mechanical vs. fly by wire. A similar regulatory regime already exists in aircraft design, which could be a guiding example.

jjk166 6y ago

That's the opposite of reality. More mechanical/moving parts increases risk and complexity, not reduces it. If something is misaligned or bends or breaks or even if it's just insufficiently lubricated, you're SOL. Solid state transistors are stupid reliable in comparison. Further, implementing more than two or three levels of conditional logic with mechanisms is prohibitively expensive, whereas you can have thousands of checks in a digital system practically for free. Finally, a mechanical system can only handle situations it is designed for, upgrades are difficult and sometimes impossible to implement. On the other hand digital logic is easily updated to deal with unforeseen issues. Electricity is just as much a physical phenomenon as fluid flow or mechanical stress. [Edit] I'm a mechanical engineer, no one should be any more enthusiastic about doing things mechanically than I am.

TheCapeGreek 6y ago
Sure, because taking out the digital system to repair/upgrade it (unless your tap is somehow wireless too) is totally different from taking out these damaged/upgradeable mechanical mechanisms.
- jjk166 6y ago
  If you don't see a difference between updating software and hardware, I don't really know what to say.
umvi 6y ago
I guess the point is... say you use a microchip connected to a sensor to determine when to switch off a gas nozzle. It would work, of course, but if the sensor fails, or there's a bug in the firmware, or if power fails... gas spills everywhere. Whereas if you couple the trigger to a physical phenomenon that doesn't need electricity to work, now there's only one thing that can fail, and that's the part converting the physical phenomenon to a mechanical action. I love computers as much as anyone on this site, but I like mechanical things too. I much prefer mechanical locks to digital ones. Mechanical locks are extremely reliable and well made ones work for centuries. Whereas I wouldn't trust a modern fingerprint padlock at all - what if the sensor fails? What if the batteries run out? What if moisture gets inside and corrodes the circuit board? One of the main jobs of a lock is to open when I need it to, and I just don't trust battery powered locks to do that.
- jjk166 6y ago
  > now there's only one thing that can fail That's the sticking point. People tend not to think of mechanical systems as systems full of many components any one of which can fail, but they are. In the case of the gas nozzle, you have the orifice for the airflow which can get blocked, you have the diaphragm which can break, you have the connection to the lever which can sever, you have the lever itself which can bind. Critically if it does fail, it fails catastrophically - gas spills everywhere. On the other hand consider a system employing a capacitive level sensor. When the sensor comes into contact with fluid, the dialectric constant changes and thus the voltage drop across the sensor increases to the point the valve will close itself. If anything goes wrong, say you lose power or a wire breaks or someone physically rips the nozzle off, the voltage still decreases below the cutoff threshold so the valve still closes - it fails safe. People like mechanical systems because they are easy to visualize, but that's not the same as being simple or safe. At the same time, most of our experience with electronics is with low cost consumer electronics where the price of failure is low and there is strong financial incentive to use cheap components with short life expectancies in configurations that aren't fail-safe, so we tend to associate electronics with being unreliable. But on a fundamental level, the force stopping two gears from passing through each other is exactly the same as the one producing a current when you apply a voltage to a piece of copper.