To be fair, this is not the only input scheme - but yes, to that one HN commenter: this *IS* the dominant implementation you get if you buy a cheap Chinese "standard function" calculator on Amazon today. The quirks are not ignored - they're copied when making knock-off ICs.
In the 1990s, the behavior was what you'd get from brands such as Panasonic. Casio had their own approach.
And of course, yep - there were odds and ends like RPN. But if we're talking about consumer products, immediate execution (with quirks) ruled supreme.
HP RPN calculators (which your articles and your collection are too light on, in my opinion) have their own set of UI problems. In general, the behavior is defined more precisely. But some questions remain ("do I need to press ENTER now or can I start entering a number?", "if I press ENTER now, will I confirm the currently displayed number into register X, or will I also get a copy in register Y?") and the answers aren't always obvious. The answers might also be different depending on which generation of calculators you are looking at.
Yeah. I think you're right to call me out on that. I didn't cover RPN in any detail because it was basically an HP thing, and it was aimed squarely at business and scientific users. The earlier history article also doesn't really dwell on input methods. But now that I've done this article, a followup on RPN would make sense...
I always had a soft spot for the HP 12C with "Reverse Polish" notation. I had no idea why these existed after Excel, but basically you needed it for the CFA exam. That is the exam for guys lacking the interest or math skills to go actuarial but too smart to be associated with lowly CPA's. I made that up. All three exams are for different career paths, but that calculator is still wild. You enter the numbers, then the function. It had a handful of finance functions (IRR, NPV,...). Now, It is not that hard to program such functions in a graphing calculator, but large numbers of finance types have an unreasonable fear of programming (a full blown phobia in many).
That said, reading this makes me feel less nerdy in owning half a dozen calculators. If the end of the world came, I think the TI-92 would be my go-to. In a perfect world I would find my old Sharp calculator. That was x86-powered and sexy. It was my first computer-like thing. I swear it exists somewhere.
This is mostly just a consiquence of math notation being complex. Some calculators had the user enter formula's using schemes like RPN, where numbers pushed onto a stack beforehand and operations can be executed immediately. This has a few advantages, like removing the need to write down and re-enter numbers during long calculations, and allowing you to sanity check all the numbers before doing any math.
Of course, in the end, intuitiveness won out (no one wants to learn how to use a calculator) and more complex calculations are just done on a computer. The stuff that mattered to nerds wasn't all that important to most people.
But, if you run a Unix based system (mostly Linux and Mac OS), try running:
To be fair, this is not the only input scheme - but yes, to that one HN commenter: this *IS* the dominant implementation you get if you buy a cheap Chinese "standard function" calculator on Amazon today. The quirks are not ignored - they're copied when making knock-off ICs.
In the 1990s, the behavior was what you'd get from brands such as Panasonic. Casio had their own approach.
And of course, yep - there were odds and ends like RPN. But if we're talking about consumer products, immediate execution (with quirks) ruled supreme.
This is where an RPN scheme makes the implementation much simpler.
HP RPN calculators (which your articles and your collection are too light on, in my opinion) have their own set of UI problems. In general, the behavior is defined more precisely. But some questions remain ("do I need to press ENTER now or can I start entering a number?", "if I press ENTER now, will I confirm the currently displayed number into register X, or will I also get a copy in register Y?") and the answers aren't always obvious. The answers might also be different depending on which generation of calculators you are looking at.
It depends a lot on what generation. I have a 42S and a 48GX. The 42 has confusing semantics, but the 48 is completely consistent and makes sense.
Yeah. I think you're right to call me out on that. I didn't cover RPN in any detail because it was basically an HP thing, and it was aimed squarely at business and scientific users. The earlier history article also doesn't really dwell on input methods. But now that I've done this article, a followup on RPN would make sense...
I, for one, would be rather happy to read something about RPN, so that I can vehemently disagree with your take on it.
I always had a soft spot for the HP 12C with "Reverse Polish" notation. I had no idea why these existed after Excel, but basically you needed it for the CFA exam. That is the exam for guys lacking the interest or math skills to go actuarial but too smart to be associated with lowly CPA's. I made that up. All three exams are for different career paths, but that calculator is still wild. You enter the numbers, then the function. It had a handful of finance functions (IRR, NPV,...). Now, It is not that hard to program such functions in a graphing calculator, but large numbers of finance types have an unreasonable fear of programming (a full blown phobia in many).
That said, reading this makes me feel less nerdy in owning half a dozen calculators. If the end of the world came, I think the TI-92 would be my go-to. In a perfect world I would find my old Sharp calculator. That was x86-powered and sexy. It was my first computer-like thing. I swear it exists somewhere.
This is mostly just a consiquence of math notation being complex. Some calculators had the user enter formula's using schemes like RPN, where numbers pushed onto a stack beforehand and operations can be executed immediately. This has a few advantages, like removing the need to write down and re-enter numbers during long calculations, and allowing you to sanity check all the numbers before doing any math.
Of course, in the end, intuitiveness won out (no one wants to learn how to use a calculator) and more complex calculations are just done on a computer. The stuff that mattered to nerds wasn't all that important to most people.
But, if you run a Unix based system (mostly Linux and Mac OS), try running:
dc -e '1 2 + p'
odds are, you will get 3.