The Future of RPN Calculators

Noksagt writes "HP's recent release of the 33s won't be the last RPN calculator. Former HPers at Hydrix are hyping an impressive Linux-based PDA/calculator, named Qonos. They have a survey up regarding features, etc. More information can be found at comp.sys.hp48 or The Museum of HP Calculators. A new open hardware project called OpenRPN has also begun. Their mission is to produce horizontal and vertical format scientific RPN calcs and later a graphing calc."

Comments

Re:My survey response

[DonServo]

Inconvenient? I can perform complex calculations MUCH faster using RPN than I can with an algebraic calculator. No worries about misplaced parantheses!

I still have an RPN calculator

[Marxist Hacker 42]

I installed The DIV Calculator for PocketPC on my Windows CE device because I missed my old HP32s.

Re:My survey response

[Short Circuit]

Spoken like someone who's never used it.

I was addicted after using it for, oh, a month. It took some getting used to, though. Unfortunately, I didn't realize I liked it until the calculator (HP 48g+) was stolen. I pity the poor sap here on campus who buys it from whoever stole it, but isn't able to use it.

Zaurus Support?

[tji]

I see from the information about the "Qonos" Linux based calc that it uses an XScale (ARM) processor - like the Zaurus's do.

If it's using the same OS, on very similar hardware, the Zaurus would be a good development/prototype platform. And, it would give us Zaurus owners a good calculator option.

??? - More Detail, please.

[students]

I am going to start calculus next year as a softmore. Can you explain this in more detail, with out actually using calculus? I want to know because I would like to buy a new calculator this summer, and I want to share this with my calculus teacher. I've tried RPN on my TI-83 plus (with RMN III and CABAMAP applications), and found it to be minorly surperior.

Re:??? - More Detail, please.

[shobadobs]

I would suggest getting the HP-49G+. (I am a 49G user.) RPN might only be "minorly better," but the real question comes down to the operating system's user interfaces. The 49G's is simply far better. In RPN mode, the user can use all the calculator's features without having to jump through dialogue boxes all the time. For example, say I wanted to graph a sine curve. Instead of having to go to the "Y=" menu and such, with which we're all familiar, I merely need to type X [SIN] EQ [STO] [PLOT] [F4] [F5] [F6]. (and I can press F4 F5 F6 in very quick succession) It is really nice to use.

After having used a TI-89, then an HP-49G, when I try going back to the TI-89 for some purpose (maybe a game, or a program i once made), I really start to notice how annoying and slow the 89's UI is.

The HP is also much more customizable. The ability to remap the keyboard is quick, easy, and built in. There are 128 system flags for changing user settings, accessible via the MODES menu. The HP has no limit to the depth of the directory tree. Its menus are better -- they are at the bottom of the screen, somewhat like the TI-85 and Ti-86's menus, rather than having to type Shift MATH 4 4 every time you need to use a specific function.

One of the neatest things is programming. Its programming language, Reverse Polish Lisp, is simple yet powerful. Plus, it has a built-in compiler for System Reverse Polish Lisp, and ASM.

Plus, text-editing on the HP 49G/G+ is much, much easier. When in ALPHA mode, one can type both letters and numbers because the number pad has no letters on it.

I used it for BC Calculus this year, and it easily served my needs.

Re:??? - More Detail, please.

[DrPascal]

Certainly not trying to argue here, but I think your example is just one of habit more than efficiency.

[Y=][CLEAR][SIN][X][GRAPH]

Isn't this five keys, vs eight? I can't talk about the other things you mentioned (which might be better), but the graphing example wasn't helping your case much.

Re:??? - More Detail, please.

[Anonymous Coward]

I am going to start calculus next year as a softmore.

In college? You should've started it as a Freshman. Calculus is extraordinarily important. The ideal situation is for a person to have an Engineer father and a Physicist mother (or Physicist father and Engineer mother) and learn theoretical and applied Calculus at around age 4. Most of us didn't have that, so we make up for lost time.

I would like to buy a new calculator this summer

Buy a TI-89. They were $150 when I bought mine four (or so) years ago, and yours will cost $150. The TI-89 is a good deal faster than all but one or two models of HP calculators, and the function interface is powerful and intuitive. You can perform indefinite integrals to a great degree using symbolics, and just evaluate definite integrals. The TI-89 also allows you to select previous function entries. If you screw up one parenthesis entering a huge calculation (which will all be one one line), just select the long calculation out of history and correct your mistake. With the HP, you'll get to reenter the whole thing, and you end up not being able to see what you entered.

Ideally, you shouldn't use a calculator at all. Learning math is like learning to do anything else--expect to screw stuff up until you get the hang of it. Using a calculator will extend the screwing-stuff-up period. It's having such a negative effect that colleges are starting to ban calculators in the math department (which doesn't mean you shouldn't buy the TI-89 anyway). Learning the material is the first priority, but that doesn't mean getting a decent grade _isn't_ important. Nobody wants to spend the next couple of semesters worrying if their C in calculus will lose them a scholarship. Grades are a source of stress that you can completely eliminate by working hard.

The chain rule deals with functions. A function is a computation on a variable. If I define f(x) to be "2x + 3", then f(3) -> 9, and f(4) -> 11, etc. I don't always have to use a variable as the input. I could use another function: define g(x) = 3x. Then I come up with fun mixtures like f(g(3)), and so on.

Calculus has a concept called the ``derivative.'' The derivative of a function is another function that tells you the slope of the first function at that location (note how this is extremely compressed). When I say f(x) = 2x+3, you should think of that on a graph, with x running along the independent axis and the f(x) along the dependent axis. That particular function is a line with a y-intercept of 3 and a slope of 2. The derivative of the function is 2.

The chain rule gives you a way to calculate the derivative of a complex function. Define z(x) = f(g(x)) from above. You calculate the derivative of z(x) by using the chain rule. In this case, d(z(x))/dx = d(f)(g(x))/dx * d(g(x))/dx. The thing to note here is how cool it'd be to have a LaTeX interpretter so that this text stuff didn't look so much like ass.

Re:??? - More Detail, please.

[TastyWords]

Big hint: even if calculators are permitted, do your best to see if you can do it long-hand. If you punch everything in, you aren't actually doing the work and you aren't actually learning the process, no matter what someone says ("The only way you can punch the right keys is if you understand what you're doing.") longhand provides better comprehension of the problem. Besides, if you get an answer you don't trust, it's easier to look through what you've done and trace your work. Punch it into a calculator and you may or may not have gotten it right - if nothing else, use it for verification if you are so inclined. But be careful - I had a prof in Advanced Calc and DiffEq who loved to give problems whose answer would be "2" or "3.7" - something which would instill doubt in most minds - "How can it have such a simple result?"

There is a real-world situation for this: to learn to walk a tight-rope, bring it down until it's 6" off the floor. Most people have no problems practicing that way. But it's the same rope. The only difference is in your mind.

The only "real math" classes I took in college were Calc III, Advanced Calc, and DiffEq. I hated real math that much. Having studied under Hofstadter and Erdos while in high school made me realize I needed to be open & free in my thinking - abstract algebra, group theory, and Galois' work made me realize how much more suited I was for "pure math". It's so open and you can create anything you want to with your imagination.
I just wish the regular high school classes taught students something other than "real math" algebra and calculus so students could see a difference and which is better suited for them.

Re:??? - More Detail, please.

[students]

Wow, you brag as much as I do. I take the practical mathematics classes you are complaining about, and I still know that abstract work is more suited to me. But I have Stephen Hawking and Andrew Wiles to thank, not my teachers.

Shameless Plug

[kwiqsilver]

For those who love RPN, check out this RPN calculator for GTK (now GTKmm2.2 compliant):
ghsiloP
I discovered last night that the stack doesn't scroll when you add lots of values, but I'll fix that soon.

Re:My $2E-2

[Noehre]

Amen, I used a 32S all through middle school and into high school before it gave up and died on me. It was subsequently replaced by a 48GX which I believe I lost recently (I can't find it).

If they made a 32S with 4 visible lines of stack, I could die and go to calculator heaven. Best calculator ever

Re:My survey response

[khuber]

There's a huge problem with regular RPN and complex equations - you can't go back and edit an expression you just evaluated like you can with algebraic.

Having to reenter a large formula just to change a constant is a pain with RPN. Once you get into that Mathematica/Maple working mode where you edit expressions I think algebraic wins.

I have an HP48SX and HP48GX which I like but never use -- I use a great command line program called "calc" in Linux. If I don't have calc I use bc. Both have readline support (assuming GNU bc).

RPN is cool for more than engineering

[TykeClone]

I've got an old HP28S and work for a bank

During "Y2K", we had to do lots of testing of interest accruals, and it was a lifesaver.

LONG LIVE RPN.

Re:What's wrong with just being a calculator?

[exp(pi*sqrt(163))]

It's a calculator if you just need to press one button to find the cos of a value.

Re:Natural it isn't

[pdbogen]

Hell, I first used it in the mid-nineties, and I'm near-religious. Infix is so slow when you're doing simple (complex) arithmatic... You mean I have to use parentheses? WTF?

People can learn RPN easily

[AceyMan]

My wife (then live-in, who is *not* a mathy) adapted easily to RPN for everyday math (checkbook, etc).

When I explained to her, "Hey, its the only kind of calculator I have", she sat with me for 5 minutes and picked it right up. It is not hard *if* people will just open their brain for a moment.

Didn't our moms teach us to try new things?

-> proud 15C, 12C, and 32SII owner

The past and future of RPN

[Nybble's Byte]

I've been using HP calculators ever since buying an HP-35 in 1973. I wandered off the path a couple of times but not for long, because RPN just makes total sense.

The tremendous success of the HP-12C in business is proof enough. And ask any HP-41 user how he likes his machine. The thing is a tool that helps you get a job done in as few steps as possible.

I use a 48GX or 49g+ every day. I much prefer the 48GX keyboard and the location of the Enter key (although I'm slowly getting used to that), but everything else about the 49g+ is better: much faster, much better and slightly larger display, and lots more 'stuff'.

Unlike some people I don't mind the 48/49 implementation of RPN (actually RPL) compared with the old style 4-level stack, but a lot of old timers refuse to use the RPL machines that started with the 28. At the same time I still like the 41/42 a lot. There's surely a place for machines like the 32S II. It remains to be seen how the 33S with its odd looking keyboard does in the marketplace.

I'm one of a very few who have an HP PDA based on Linux that never made it to production. Display contrast isn't very good, but otherwise it's decent but not feature laden because the project got killed and people lost their jobs. For this reason it's heartening that someone's talking about doing an HP handheld running Linux, and I'll support their efforts any way I can. It's about time for the user community to pitch in - this reminds me of the PPC 44 project talked about what, 20 years ago?

Re:this needs linux?

[n3k5]

Everyone in the Slashdot community seems so worried about the Microsoft monoculture, and yet here we are pushing linux into every possible niche. It can't be optimal to have the same OS running on both our calculators and our supercomputers.

I surely wouldn't buy an expensive calculator that runs out of batteries much faster than my PDA, but I'm glad that people are doing research in that area and maybe improve the system so it becomes easier on the hardware and cheaper; just as I think it's a good thing for Linux to be adapted fpor supercomputers, even though I won't ever have one. Imagine everyone would regard Linux as being strictly 'for PCs only' and then some company makes a Windows (CE) based calculator, and Microsoft says, "See? You can't use Linux for that!" I'd rather not grant them that satisfaction ;-)

Re:The future of RPN calculators...

[exp(pi*sqrt(163))]

It's pretty common English usage to use tautologies to stress a point. In this case someone was under the misguided impression that two things were different. It's common in such a situation to apply stress when you correct the original speaker so as to draw attention to that part of your sentence that is most significant. In this case, same exact, performs this function quite well.

These kinds of subtleties can be quite hard for non-native English speakers to discern but you'll figure it out eventually.

Re:12 C replacement

[Eric Smith]

Why recommend the older when the newer is much faster with more memory?

And more bugs, too. For instance, when the program is longer than 260 lines, many GTO instructions that should be valid give an error 4. So the extra program memory isn't all that useful, unless you're writing mostly straight-line code. See the November/December 2003 issue of Datafile (the HPCC club magazine) for a list of more HP 12C Platinum bugs, most of which admittedly not as serious as the GTO problem.

Hewlett-Packard used to put a huge amount of effort into making sure that there were no obvious bugs. And when a nasty bug did slip through, they usually would provide ROM upgrades on request. But Hewlett-Packard no longer exists; it's been replaced by the new HP, which has abandoned the "HP Way" for what is apparently the Compaq Way. There's no word of any ROM upgrade for the HP 12C Platinum.

The existence of serious bugs in something as simple as the GTO instruction makes me rather skeptical of the accuracy of the financial calculations. Some of the financial calculations such as solving for interest rates are actually quite challenging to do accurately, and Hewlett-Packard spent a lot of time on numerical analysis to make sure that they did a good job of it. I rather doubt that the new HP has done nearly as much homework on the mathematical algorithms the HP 12C Platinum.

The original HP-12C also has a better keyboard. Hewlett-Packard invested more effort into making the calculators ergonomic. When you buy a $100 calculator, you should be able to get reliable keys with good tactile feedback and double-shot injection molding so that the legends don't wear off. But HP doesn't care about this. What they've forgotten is that by cutting corners and making the products as cheap as possible (but not any more inexpensive), they are not only failing to cultivate new customer loyalty, but are losing the loyalty of their existing customer base.

I have yet to meet someone that actually needs a faster 12C, or one with more memory. But I don't doubt that there are a few people somewhere that do. For everyone else, I strongly recommend the HP-12C over the HP 12C Platinum.

RPN Heaven? You should try japanese...

[Anonymous Coward]

If you want the real RPN, you should try learning japanese. In Japan, the verb(=function) is always after the subject and object. Sounds pretty much like RPN to me ;)

p.e.
Nihongo wa omoshiroi desu.
(Japanese Language, convenient) is

there you go!