I fell in love with RPN way back in the HP heyday. Good stuff.
Recommend a good RPN Android app, please!
Exactly! As I usually quip: when you’re adding numbers on paper, do you write down the first number, add, and only then write down the second number?
What “algebraic” afficionados fail to understand is that the “equals” key actually does the addition. The “plus” key only indicates your intention to perform addition some time in the future.
I still miss my HP-41C.
But I do have an RPN app for my smartphone. Of course, I rarely have a need to do anything more complex than simple math these days, but if I do I know where to go (assuming I can still remember how RPN works after all these years 
)
I don’t miss my 41. Not that I didn’t love it. I entered college with an HP-25 and graduated with an HP-41. I was heavily into the 41, synthetic programming, etc. but as soon as the Voyager series came out, I was in love. It was just so small. Truly shirt pocked sized but fully scientific and programmable. I picked up a 15C in 1982 and it’s still with me today. I believe I’ve only changed the batteries twice, it’s so stingy with the juice. I’ve also owned a 16C (programmers) and a 12C (business) which is still in production after more than thirty years!
(hefting the 8.5 ounce TI-59 with a restored battery pack I’ve been schlepping to work because I have a cheapass phone with no camera and only a 4-function calculator that has no funk factor at all but for security purposes that are stricter than the NSA I’d get fired slightly faster for possessing a phone or especially a camera on the sales floor than I would for using a calculator and wondering why I never restored the TI-58C I found in a box because it has to be lighter without that stupid and IIRC broken card reader) You might have a point.
With your next one, and now that you have the bee in your bonnet I know you will get one, you restore the battery pack with three tabbed AA ni-cads. Regular ni-cads don’t fit. And if you need to replace a crapped-out charger, they usually fail when the cable at the wall wart breaks. Solder the cable to a power supply from a retired cell phone (preferably 5V but 3V works) and you will be the envy of a nerdly minority of your co-workers.
I mean, DEATH TO ALGEBRAIC NOTATION! LONG LIVE RPM! OR REM! OR WHATEVER THEY CALL IT!
Yeah, that’s the one we had!
http://www.hpmuseum.org/hp9100.htm
(Scroll down toward the bottom for the pictures.)
The one they show as the 9100-A looks most like what I remember we had. We didn’t have the printer unit.
And to this day, I still have, and sometimes use, a HP 32S. I used it a lot, especially, when I took a Statistics class about 20 years ago. It has built-in bivariate statistical function (which I didn’t use much), but I did write my own program to compute areas under standard bell curves.
I also used it for some numeric integration in Calculus class. We had a take-home final that included a line integral problem that could only be done with numeric integration. (I think the point of the problem was to see how many students would realize that.) I wrote three programs to do it three different ways, using trapezoidal and parabolic integration and a primitive way that just broke curve into pieces and estimated each segment length with the distance formula. I compare how many terms or iterations it took to get a given accuracy with each method and documented all that. (Waaaaaaay beyond what the problem actually asked for.) The prof remarked that my final exam looked more like a dissertation.
Hope you got an A+ Senegoid on the exam.
I had a TI-51 before my HP 15C. Since it was programmable, the very first problem I came up with was to decide what an alternating harmonic series (i. e., 1 - 1/2 + 1/3 - 1/4 + 1/5 + …) converges to. It took a few minutes, but ended up being the natural log of two (0.6931…). I leave to the rest of y’all to figure out why that’s the case.
The way I’d explain it:
Let F(x) = 1 - x + x^2 - x^3 + x^4 - … Note that F(x) = 1 - x * F(x); thus, F(x) = 1/(1 + x).
I.e., (1 - x + x^2 - x^3 + x^4 - …) = 1/(1 + x). Integrating both sides from 0 to 1, we get 1 - 1/2 + 1/3 - 1/4 + 1/5 - … = ln(2). [With some use of Abel’s theorem if you are being rather pedantic]
Personally, I use this one. Love it.
ETA: On my PC, I either use XCalc for its great usability, or the HP-15C simulator for its nostalgia value
And I still haven’t gotten over the fact that my trusty HP-15C - which I bought the during my first term in college, for about 10% of my student loan for that term - was stolen when I was a PhD student :mad:
D’oh!
Flashback to my first year in college: that’s exactly how the math profs solved when I showed up to reveal my “discovery”.
Kudos to you.
Hey, kudos to you as well for discovering it yourself.
Sorry to hear about that…:(:mad:
I count my HP15C and my SVEA 123 campstove as two of my most treasured possessions; because, they represent the best times of my life. (I got the stove as a child; it’s now 43 years old and still just as cantankerous to light up, but roars like a jet engine when it get going).
It’s a funny thing about this thread–there’s a real joy talking about using the very first wave of a portable digital technology for students that only hinted of the things to come.
do they make any other kind? I only use RPN calculators. If it isn’t RPN, it essentially doesn’t work (for me). If there are other kinds, I have never run across them.
I don’t know if postfix is necessarily simpler, but it’s sure easier to implement with a stack, and pretty thrifty to boot.
I’d almost guarantee those are the reasons that the old 1960s era HP calculators are RPN/postfix.
You’re right about it just being a matter of teaching, but I’m still not convinced that postfix notation is particularly intuitive.
What about prefix notation, as in f(x, y, z, …)? Is that particularly intuitive?
Really, we should be writing these things out as trees rather than linear strings; alas…
Ditto. I took to RPN very quickly when I was in high school, and, ever since then, conventional format calculators feel extremely clumsy to me when I use them.
I’m very glad that, should my beloved HP 11C ever die, there are some RPN products you can still purchase. Also that the RPN app on my phone works great!
Viva RPN!
Who said that? I thought that it was RPN calculators that were considered obsolete.
I have two calculators (besides ones on my phone and iPod Touch): an HP 48GX (a replacement for a 28C which suffered from the common problem of the battery cover snapping off when I tried to change the batteries), and a cheap (less than $10, IIRC) Samsung scientific calculator. Before that, my first two calculators were a 1970s-era Radio Shack scientific, and a TI-59 (and yes, I had the printer as well).
Note to self: once you get Eclipse Kepler running, try designing a scientific calculator app of your own…
Well, if you start with a tree, and then flatten it out into a single line, you end up with something like Lisp.
Preach it!
To the OP, how did you ever grok recursion if you didn’t grok postfix?
Or standard function notation, or Reverse Polish Notation, or whatever. It’s just a matter of how you choose to write “The tree with root so-and-so and children such-and-such”; you can flatten trees preorder or postorder, as you like, and you can use whatever symbols (parentheses, commas, etc.) you like in that process.
[The one non-arbitrary design decision is the business of whether the label on an internal node determines its number of children (thus completely obviating the need for bracketing the list of children) or not.]
Well, I don’t think recursion is any sense innately connected to postfix notation; semantics and notation are orthogonal.
Yes, postfix notation in some sense corresponds to the control flow of stack-based evaluation… but then, recursion isn’t innately tied to stack-based evaluation either. Many a programmer has used a programming language with recursion without the slightest idea about a stack; for that matter, a programming language with recursion can just as well be implemented with some means other than stacks, and this might even be more appropriate to the goals of the language implementation. Particularly if one is interested in exploiting the level of parallelism available in independent evaluation of a function’s separate arguments, the stack model will not be a good way of thinking about a recursion [rather, one will want to think only of the tree representing function calls and their dependencies].
(An interesting observation is that, in a language that already has built-in support for recursion, it’s actually somewhat easier to write a parser for prefix expressions than for postfix expressions)
