The R47: A new physical RPN calculator
swissmicros.com
swissmicros.com
The first is Termux, which provides a gnu userspace atop the Android kernel. This app is pretty old, and well-tested. There is an active and helpful Termux community. But it has some downsides: 1) The version of Termux in the Google Play Store is not the preferred and maintained version, although the Play Store version does work. The preferred version is in F-Droid, but the future of F-Droid itself is uncertain in the light of recent Google policy decisions. 2) Termux does not have access to directories such as /proc, /sys etc, which prevents some gnu/Linux utilities from working and 3) The Termux filesystem layout is very non-standard, so unless a program has been packaged explicitly for Termux, installation will probably be messy. I was able to get most, but not all of the Python packages I use frequently, to run within Termux. I could not get astropy to work, for example. Termux has nice usability features like pinch-to-zoom to change the font size. Termux requests a wakelock, and if you grant the wakelock then the OS will not throttle the app when your phone is locked.
The other option is the relatively recently added "terminal" app. terminal runs a plain-vanilla Debian Linux OS within a VM. Its file system is laid out exactly as you would expect, so if you want to get iPython and lots of libraries, you can just run the Anaconda Python installation script, and it will run unmodified with no errors. Nice! You can also install other nice desktop-style apps like VeraCrypt. There are a few downsides: 1) The OS will throttle the app, and occasionally kill the app, when the app is not actively being used interactively. 2) I have found no way to change the tiny font. 3) It's a Google app, so it might disappear for no good reason, as so many Google products do.
Both of these options work especially nicely on a foldable phone, because then the tiny phone keyboard is much less of an issue. A foldable phone plus the terminal app really is a pocket Linux computer.
I've never used a Samsung phone, but I think their DeX environment might allow you to do the same things that the "terminal" app supplies.
If you can't get either of the above to work, give Termux a try. It's not full gnu/Linux, but it's pretty close.
Of course, then a $5 calculator may serve you better ...
There are just some user interface advantages for me in having a physical calculator. The only thing stopping me from buying this is the price tag.
Mathematica does everything your python interpreter does and more — but that doesn't mean python becomes useless because of that. It's good for certain things, while Mathematica is good for other things.
Similarly with calculators (especially HP calculators), which have become a bit of a lost art (an elegant weapon for a more civilized age). I use them all the time when doing any kind of engineering or hobby work. Good luck using your computer with python one-handed in a workshop next to a CNC.
One thing I do agree with is that plotting functions or working with larger matrices on calculators makes little sense.
https://en.wikipedia.org/wiki/Maxima_(software)
Cayley (1975–1993), now Magma, predates it somewhat and has recently been used to crack post Quantum cryptography candidates.
https://dl.acm.org/doi/10.1145/1089355.1089359
https://en.wikipedia.org/wiki/Magma_(computer_algebra_system...
(Maybe ggp is thinking of the Matlab clone Octave, which is?)
In high school, my friends got onto the Numworks bandwagon, and we even used them on the SATs and AP tests (they were explicitly allowed). To be fair, this is before Numworks locked down their calculators and the alternative firmwares (Omega) died off, but maybe there are jailbreaks now and things are as they were before.
The difference (as far as I know) is that Numworks (and Casio, etc.) are mainly used by students, while SwissMicros (and HP) are used for very quick problem-solving on the job.
There was a time when RPN was waaay better than the TI and other competition, but in my mind that time is now past us. I can use both designs well and will take a TI-89 over a Swiss Micros (mind you I own several) almost any day as it is more straightforward. That's my opinion anyway. This improvement over the DM42 is pretty cool though.
Another Swiss Micros use is for people who work in a lab with gloves and can use a calculator better than a phone.
So I doubt I'll buy this one, even though I'm happy someone made it.
The R47 has been many years in the making and is a small open source project which has collaborated with the Swiss manufacturer of calculators, SwissMicros. It has a superset of functions over older HP models and many more too, including complex solve, default 34 digit decimal precision, 1000 digit integers, graphing, extensive complex support, etc and is substantially customisable.
I have no affiliation with the project, but excited that there is a new RPN machine commercially available.
The original calculators, from the discrete HP9100A onwards, pushed tech to its limits.
The HP65 (1975) was a jaw-dropping masterpiece. When most calculators were four function, and scientific calculators were still exotic, a pocket-sized programmable calculator with a magnetic card reader was beyond the imagination of most engineers, never mind most users.
This is more of a nostalgic tribute act. It's nice it exists. But it's looking backwards, not forwards.
Don't rely on your notes come quiz time, if you can get away with it, but by all means do take them during the learning process. And in the real world, you're allowed to refer back to your notes as often as you want to.
I like solving math puzzles, and I'm often realizing that I'm missing something I've forgotten since middle school or high school - some formula that I know exists that would make a problem trivial. I'd like to re-learn that stuff, especially as my kiddo starts advancing in math classes.
(Also, reinforced what I remembered from Trig, I finally fully grokked it. Except for trig identities, ha ha.)
The book was targeting DOS and C++. I wrote it in C for the Macintosh. (That required that I figured out what was going on.)
Go write one in JavaScript using an HTML5 canvas as your buffer. (I made a quick pass at doing something similar [2] — but do it on your own, don't follow my link below or look for the sources for Phosphor3D on GitHub.)
[1] https://archive.org/details/build-your-own-flight-sim-in-c-d...
Math is a broad subject. This is something LLMs are actually reasonably good at: ask them for textbook recommendations, and get into a dialogue about which sub-areas of math you're interested in and what level you're currently at, whether you want pure math (more theorem-proof focused) or applied math (more practice solving concrete problems, e.g. finding lots of derivatives and integrals). Toss in names of books that have been recommended and ask where they fit in to the LLM's other recommendations.
LLMs don't understand the math, but they're trained on a lot of discussions and recommendations for math books, and have a reasonably good sense of what level different books are at.
Download multiple recommendations in each area and try them all out. Seeing how different authors start out approaching largely the same material will help you conceptualize it better than just relying on a single approach. There's no universal "right" book to learn from. I wouldn't buy non-free textbooks without trying them out first.
Youtube has a lot of math lecture series, which can help if you're stuck on a particular point, but they're not the same as doing problem sets yourself.
I'm also referring to the faster models, not the slow and expensive deep thinking ones which I have little experience with. I don't see how reasoning would enable deep thinking models to meaningfully evaluate textbook pedagogy, either.
They DO understand what they are doing. When I ask it to solve math problems, it goes through the several (many) steps involved (e.g. e.g. "apply the chain rule" while doing partial differentiation on a term in a Jacobian matrix). It gets pretty tedious when solving systems of linear equations, where it goes through each step of the Gauss-Jordan elimination while doing an LU decomposition, row by row. But one learns to ignore the blah-blah. Step by step, in absolutely ridiculous detail. The point: they absolutely 100% understand what they are doing, and understand it in minute detail.
It's clearly NOT regurgitating something that it has literally seen before, because the level of detail is beyond ridiculous for a human. It is applying generalized rules to specific concrete problems, and doing so with some level of strategic thinking.
Where did it learn those generalized principles, and how did it learn to do that? With absolute certainty, there are math textbooks among the materials they have been trained on. And they certainly learned it from SOMEWHERE. Probably math textbooks. How did they learn to generalize and think strategically? Well, that's the big mystery, isn't it? But they do.
The very best models achieve high scores on Math Olympiad problem sets (so competitive with some of the best minds on the planet). And Terrence Tau (greatest living mathematician) declares state-of-the-art models to be "better than most of my post-graduate students".
And what they can and cannot do is increasing by leaps and bounds on a weekly or monthly basis right now. It's hard to keep up. I frequently find that they can do things this week, that they could not do a week or a month ago. Startling, and quite utterly amazing.
Most of the time, I am using Claude Sonnet 4.5 as my coding agent, for which I pay $10/month. Measured IQ of 110, I think, with an IQ of 120 if you flip it into thinking mode. But only because there isn't enough undergraduate level mathematics in a standard IQ test. Claude Sonnet 4.5 is also available for free here: https://claude.ai/chats (during periods of heavy load, it may fall back to simpler models). I often use the free web interface instead of the Coding Agent interface for math problems, because it's easier to read mathematical equations in the browser version. version). And I generally use the free version of Claude instead of Google Search these days.
My experience with people who have LLM subscriptions of any kind is that they use them all the time and would immediately ask an LLM that kind of question, rather than asking on a web forum that's not even dedicated to math. So I think it's a fair presumption that someone asking that question doesn't have access to the best commercial models.
On the largely irrelevant question of what math LLMs can do, although Opus may do better, Sonnet can follow procedures sometimes but not consistently. It has blind spots and can't scale procedures; beyond certain numbers or dimensions or problem complexity, it just guesses (wrong). And those limits are quite low. If you want 2 simple examples:
4294967297*1331
Invert this matrix: m=[1 0 5 0 3 7; 2 3 0 3 3 2; 1 0 1 1 0 1; 3 5 3 5 1 2; 2 4 3 2 1 5; 1 0 5 2 1 5]
LLMs follow procedures, but whimsically. Better LLMs will be less whimsical, but they still won't be fully competent unless they digest questions into more formal terms and then interface with an engine like Wolfram.
It was some course on edx [2]. Can't find it right now, but you might find another course which uses it.
But it's unlikely that I buy one because (1) I tend to be close to computers most of the time, (2) my favorite "HP calculator" is the HP9000/715 due to its sizeable HIL keyboard, its 21" CRT color screen, and its support for HP-UX, pun intended; and (3) most of the mathematics I need beyond paper, pen and blackboard can best be done in a Python Jupyter Notebook (statistics) or Mathematica (symbolic derivations).
My dad has an HP-15C that I've always thought was cool but I never really liked to use it for anything. This new calculator is also really cool but I can't imagine ever owning one. I don't need to do much calculation anymore anyway so it's not like I need any calculator at all (hence giving away my old calc). I'm also usually at my PC, or a laptop, or I can ssh to my home system with my phone (if the phone's basic calc app is insufficient). For simple calculations I've most commonly used a Python repl since like 2007, but for anything more advanced these days I'd pull out one of the PC programs I was exposed to in the later parts of college (e.g. matlab or octave+symbolic and other packages, or maxima). But I also have used Python + libraries or Lisp + libraries, and would like to someday redo my Stats education with R. Hand calculators just seem really antiquated to me now.
RPN was lovely because it was significantly faster to operate with. I remember showing up at math class having just received an HP-48G "because it's what the engineers use" and everyone else just got TI-89s because that's what the teacher had (TI gave teachers free TI-89s and did more marketing). The math teacher hosted an impromptu calculator speed contest and I beat out everyone else by a significant margin.
The shift key on my HP28C died last year. I never used the advanced features (eg the entire left-hand keypad), but loved its effectively infinite stack, and haven't been able to go back to an HP42, HP35s or the like since.
I'm now using Plus42 on my phone with the big stack option, but would love to have a physical calculator again.
A good example: Taking your FCC ham license exam does not permit use of your phone or a programmable calculator. They would have allowed me to use it.
I had practiced for the exams with my old HP 11C. It was jarring to have to switch to a TI calculator during the test.
Edit: it looks like you're right and they're all programmable. My mistake.
> Battery type 1 × CR2032 3V lithium coin cell
Very nice. Impressive.
49G aside, they all are good enough for use although probably no longer usable in some of their intended uses (re: any formal tests that would allow them?).
Every time I use for more than a couple of calculations I think how much I prefer a RPN calculator.
Just now I tried out the RPN calculator apps on the phone and I've realized that the enter key to the left above the number menu forces me to use two hands if I want to operate the calculator fast.
I think it makes more sense to stick the enter button at the bottom right near the other arithmetic operations, then I can at least operate the calculator reasonably quickly with just one hand for arithmetic.
Beyond that... do RPN calculators like these usually include the option to use infix notation?
Like many others here, I rarely use my calculator any more. My phone is just so much more powerful with a lisp REPL and python.
This is cool technology, though.
I do not need to think about my operation before I start inputting numbers. I can type in the numbers I'll need, and while seeing them I run operations on them.
1
2
+ (->3)
9
+ (->12)...no wait I meant 99
...uhhh what did I just do...9?
9
-It's just for me the most easy way to calculate things with current and power for everyday use.
And .. even the basic Windows calc app is so slow to start, I already finished until the app started.
I'll use the built in calc in macOS Spotlight for many things, but plenty of times each day I use the HP. Example: yesterday I calculated an angle in degrees by reading off approximate x, y coordinates from a graph on my screen. Swapping windows to a calculator app, repositioning them to see both at the same time, just too fiddly, and clicking the calc app moves the cursor so I have to remember the numbers as I switch context.. The physical HP in front of me leaves my screen undisturbed. Easy.
The low volume and the relatively high build quality involved pushes the prices up.
How Id love for SM to do a straight 48G with more memory and a good keyboard. (But keys are hard to do)
They've asked people to purchase a voucher on the C47/R47 site for an early-bird discount which will send some money towards the developers. This voucher can then be used to get the calculator once SwissMicros take bookings. The link is at the bottom of the Swissmicros page.
I got my DM16L two weeks ago and I love it to bits!
They even upgraded it to USB-C (even though the product description still says Mini USB), which was a pleasant surprise. I only just got into RPN calculators (and programmable calculators in the first place) a month ago, loving it so far.
I do have a few other calculators as well, so I'm definitely biased.