(12-16-2023 04:56 AM)wb.c Wrote: (12-15-2023 11:42 PM)nickapos Wrote: There is a difference between scientific calculations that require formulas, and basic number crunching you would usually do manually or with a basic calculator.
In this second case the professionals use adding machines even today and adding machines use postfix notation.
I find my Rpn/rpl calculators a lot easier to use in these cases compared to regular textbook scientific ones, but then again, unless you have been an accountant or a bank cashier as I have been many years ago you can survive with a phone calculator app just fine.
Mechanical adding machines actually use a simple two level stack with a running total. Each entry is automatically added or subtracted to/from the stack, but this is because it would be so complex to store intermediate values in purely mechanical systems. On the roll, and in the key stokes it looks like infix, but under the hood it’s postfix. Not because it was better, but because of physical limitations. The + button was actually +=. And this is why some earlier calculators had these += and -= buttons. They just carried over into electric calculators and subsequently this stack concept works quite well in data structures.
Maybe it’s just what people are used to, but the truth is RPN was made because of hardware limitations. Calculators could benefit from a the human super computer brain to simplify processing with relatively weak hardware. It literally makes you do all the work in regard to order of operations. Modern Natural/Mathprint entry modes do all the work for you. On the TIs you can even leave parentheses multiplication operators out and it will add them in automatically for you. Because of this, you can focus on entry correctness/accuracy more than solution process and order of operations. This actually speeds things up and minimizes mistakes, which saves more time. Even with basic number crunching, say you needed to add 10 numbers together. If your total was off, you’d have no way to know where you made your mistake. You could redo the calculation again, but you risk another mistake. With modern input, I can see each operand and operator and modify them if needed. For more complex number crunching in RPN you need to spend time thinking about order of operations and how to approach the problem to make good use of your stack. Whereas on modern input, I can input my problem exactly how it is written on paper once, and get a solution without spending a second in order of operations, or thinking about my stack. If I made a mistake on my input, I can see it and fix it and get an updated answer quickly, but in RPN I would need to redo the entire problem if my final output was wrong.
Again, I want the emphasize modern “natural” entry as Casio calls it and “MathPrint” as TI calls it, that allow you to enter a problem as shown on paper. I’m not taking about single line algebraic entry. Sadly a lot of advanced graphing calculators don’t have this entry style on dedicated keys and require entering submenus to access these entry styles. Of the TIs for example, only the French TI-83 premium CE has a dedicated MathPrint fraction entry key. Most scientific calculators have this, both in Casio and TI, it is common place. For me, this is what broke RPN. No matter how good the DM42 is, it will never afford me the speed, ease, and flexibility that my TI-30X Pro MathPrint gives me, even with just number crunching.
First of all I would like to say that I mostly agree with you. Second the adding machines I used were/are electric and third, you are mistaken about the errors.
We used those machines to check all the cashier till transactions because we could find the errors. Even though the display was one line, everything was printed on the tape.
That tape was a lifesaver for us. I can not remember how many times we used to find and fix errors.
These days I simulate that behaviour using plus42 for accounting calculations and textbook style on my prime for more complex equations.