The most basic immediate execution four operation calculator might still look the same, but that’s because it’s a very simple thing and you can’t really get much wrong. For scientific calculators the UI has changed lots. As have the requirements. It used to be a specialist tool used to do thousands of calculations daily. An expensive thing that had to earn its keep. RPN and stuff like that made sense for people who could easily get back weeks of training in just a few years of being slightly more efficient while working. Now we have the natural order delayed execution thing, because the calculators are mostly for students. Who need the UI to be as easy to grasp as possible, because they won’t ever have to do enough calculations to benefit from a faster but harder UI. That doesn’t mean any of those approaches to UI is better or worse. Some things require instructions and making everything idiot-proof shouldn’t ever be the ultimate goal (check out modern computing for why!).
Because when you need to do a process a thousand times, you program it in an actual computer. Then you just have a specific interface for just your process that makes everything simple.
And the developer really only needs to understand the process for a couple months. Once it’s confirmed working correctly, you’re generally done with that piece of code.
But it has been changed a lot?
The most basic immediate execution four operation calculator might still look the same, but that’s because it’s a very simple thing and you can’t really get much wrong. For scientific calculators the UI has changed lots. As have the requirements. It used to be a specialist tool used to do thousands of calculations daily. An expensive thing that had to earn its keep. RPN and stuff like that made sense for people who could easily get back weeks of training in just a few years of being slightly more efficient while working. Now we have the natural order delayed execution thing, because the calculators are mostly for students. Who need the UI to be as easy to grasp as possible, because they won’t ever have to do enough calculations to benefit from a faster but harder UI. That doesn’t mean any of those approaches to UI is better or worse. Some things require instructions and making everything idiot-proof shouldn’t ever be the ultimate goal (check out modern computing for why!).
Because when you need to do a process a thousand times, you program it in an actual computer. Then you just have a specific interface for just your process that makes everything simple.
And the developer really only needs to understand the process for a couple months. Once it’s confirmed working correctly, you’re generally done with that piece of code.