Inspiring the next generation of RPN users

General discussion about calculators, SwissMicros or otherwise
User avatar
Walter
Posts: 3070
Joined: Tue May 02, 2017 11:13 am
Location: On a mission close to DRS, Germany

Re: Inspiring the next generation of RPN users

Post by Walter »

RPN vs. RPL or vice versa is the most religuous topic you can trigger on this forum. At the bottom line you can't convince any supporter of the other side. It seems to be a matter of habits, first contacts, sympathies, etc. ... just give up, please.
With the code memory you used implementing a stack overflow flag, you could just implement RPL instead and be done with pointless historical limitations.
Who did use any code memory for that purpose? When? Where?
WP43 SN00000, 34S, and 31S for obvious reasons; HP-35, 45, ..., 35S, 15CE, DM16L S/N# 00093, DM42β SN:00041
User avatar
dalremnei
Posts: 43
Joined: Thu Dec 10, 2020 11:32 am
Location: Scotland

Re: Inspiring the next generation of RPN users

Post by dalremnei »

The topic of this conversation is what would make sense in a school context. I think RPL would, and it's unclear why someone with no prior experience of RPN would want anything else. It's more intuitive if you're coming straight from algebraic/mathprint.

And program compatibility is irrelevant as nearly all exams ban programmable calculators.
SwissMicros DM42, DM16L, HP 12c Platinum, Prime G2 CASIO fx-9750gii, fx-991ex classwiz, fx-CG50, TEXET fx1500, TI nspire CX II-T
rprosperi
Posts: 1698
Joined: Mon Apr 24, 2017 7:48 pm
Location: New York

Re: Inspiring the next generation of RPN users

Post by rprosperi »

Qwerty? Obsolete? What do you type on? (...he asked, knowing this was exactly the point of making such a statement...)
--bob p

DM42: β00071 & 00282, DM41X: β00071 & 00656, DM10L: 071/100
User avatar
dalremnei
Posts: 43
Joined: Thu Dec 10, 2020 11:32 am
Location: Scotland

Re: Inspiring the next generation of RPN users

Post by dalremnei »

rprosperi wrote:
Sat Nov 20, 2021 9:56 pm
Qwerty? Obsolete? What do you type on? (...he asked, knowing this was exactly the point of making such a statement...)
The point of the statement was that 4STK, like QWERTY, is a standard people stick to because they're used to it, rather than technical superiority, not being asked the layout I use. I use Dvorak.
SwissMicros DM42, DM16L, HP 12c Platinum, Prime G2 CASIO fx-9750gii, fx-991ex classwiz, fx-CG50, TEXET fx1500, TI nspire CX II-T
vgoudreault
Posts: 19
Joined: Sat Oct 17, 2020 2:32 am
Location: Greater Montreal, Canada

Re: Inspiring the next generation of RPN users

Post by vgoudreault »

Regarding a stack-overflow flag: Please forget it. It will never work since there are useful situations where you want all stack levels filled. Was discussed on this forum more than once.
I fail to see how it could be an issue. The flag is set for the user to want to be told, or cleared if they do not care. Then if they overflow the stack, another flag is set, which could then be tested in a program, or set to issue a warning. Nothing else is done that the user does not specifically ask for.
But if during their casual computations, they end up with stack-overflow warns and they had set the stack depth to 4, they will know they are putting more stuff in there than they should, so either they should keep better track of their pending results, or set the stack deeper.
Peet
Posts: 256
Joined: Tue Sep 29, 2020 12:01 am
Location: Germany

Re: Inspiring the next generation of RPN users

Post by Peet »

Walter wrote:
Sat Nov 20, 2021 7:40 pm
RPN vs. RPL or vice versa is the most religuous topic you can trigger on this forum.
The classic 4 level stack still makes sense today, because almost all of the literature on RPN refers to it and can clearly better show the advantages of RPN compared to algebraic input. Plus, it's easier to understand and learn.

Entry RPN, on the other hand, has advantages e.g. for complex stack/matrix operations and data processing on the stack. But has a lack of prefilled stack, mostly no LastX and direct register access, no t-register and often an annoying error handling.

I think it makes more sense to choose the variant that better fits for the usage or the one you like better than to believe in technical superiority of one of these variants and to derive a religion from it.
My programmable calculators - former: CBM PR100, HP41CV, HP28S, HP11C - current: HP48G(256kB), HP35S, Prime, DM41X, DM42
Florian
Posts: 22
Joined: Mon Apr 26, 2021 1:19 pm
Location: Japan

Re: Inspiring the next generation of RPN users

Post by Florian »

I was quite undecided for a long time now whether I should also join this interesting but kind of "dangerous" discussion. But what the heck, sounds like fun. :D

So even though I am not young by my number of years I am still rather new to the world of RPN and its variants. As such I didn't have any legacy feeling when I obtained my DM42 about a year ago now. Still, I fully agree that there is no "right" and "wrong" by the personal preference for one stack type or the other. It all depends on personal use cases, personal liking and personal muscle- and brain-memory.

When I first obtained my DM42 it was, of course, a 4-level stack machine. This I had to accept as face value and I dealt with it. Then Thomas made his great move and gave us an N-level stack software and SwissMicros thankfully adapted it to their platform. For me personally that was a revelation. Suddenly the calculator felt natural to me. First from an aesthetic point of view. I only see numbers on the screen that I actually use. Additionally, having personally no need for the special nature of the T register, I was always rather annoyed by its behavior as it intended to "invade" my calculations with old numbers that I don't need any more to the point that I tried to do my calculation in three levels only. With the N-level stack I start with an empty stack and I know that at the end of the calculation I must be left with exactly one level left on the stack. If that is not the case I know that I made a mistake somewhere. That is reassuring. That is safety for me. That I required from a tool in a professional setting.

I use the calculator in my job as a physicist. In that it is a very rare case that I have an equation printed (or written) in front of me that I just need to solve. (This I can do much better, faster and safer with the PC in front of me.) Instead, the DM42 allows me to play with numbers, expressions, ways of approaching a problem. And now I can keep intermediate results visibly around to directly compare numbers. In that case I do not want to worry about running out of stack. As I don't yet know where my calculation is going to lead me I cannot pick the right position in the euqation so that I won't run out of stack. Therefore the N-level stack really opened the DM42 to me as a tool to explore physics in a very intuitive and sometimes more tangible way than my PC with Mathematica, Python and what not sometimes allows me to.

Still one feels the heritage of the DM42. Many operations do not swallow their arguments and programming in NSTK feels kind of cumbersome to me as many DROP commands are necessary to keep the stack clean. Therefore my programs mostly run in 4STK mode, allowing me quick'n'dirty approaches. Or rather programs in the way they were intended with the HP42S.

So for me I am happy that I can easily switch between the two worlds.

I would like to add that I cannot really understand the previous comments that a 4-level stack "clearly" shows the advantages of RPN or that it is easier to understand. To me it seems that understanding the idea of a stack and its use is independent on the number of available levels. Further, to me it seems that a statement such as "look how easy you can do your equations without brackets, but please be extra careful where you start your evaluation as otherwise you will be running out of stack and need to start all over" is not really that great for making a selling point to newcomers.

Oh, and where is this "repellent" disliking for RPL coming from? This seems very inappropriate and unjustified to me. Isn't that just another case of personal preference?

So I just hope that there will be many more affordable and flexible calculators out there. Calculators that give freedom of choice between RPN/RPL, between 4-levels, 8-levels and N-levels, between automatic stack lift or not, and probably between many other options that I cannot even think of now. Then everyone can make it his own machine that matches the personal needs and style best. To be inspired.
rprosperi
Posts: 1698
Joined: Mon Apr 24, 2017 7:48 pm
Location: New York

Re: Inspiring the next generation of RPN users

Post by rprosperi »

The comment about the 4-level stack 'clearly' showing it's value was referring to decades worth of easily available publications, manuals, white papers, etc. from HP and the RPN community, which show the virtues of classic RPN. It was not intended to mean 'better' than other styles, simply that there's lots of material to show it's power.

The phrase 'repellent' was coined by Valentin Albillo, a lonnng-time HP and RPN fan, as a clever phrase which uses the RPL sounds in the pejorative word. Valentin has written thousands of pages of articles, reviews, PPC/CHHU Journal articles, etc. about how to use RPN for all sorts of problem-solving and was/is a well recognized proponent for RPN in the HP community. When HP 'switched' to RPL for their high-end products, there were many ardent RPN fans that rejected the new approach as "probably really is superior for some things, but is generally not worth the added complexity, poor readability and generally arcane nature" (I am paraphrasing here, not quoting anyone).

When coming into the game with open eyes and curiosity, well after both schemes have been around, the near-religious fervor and even name-calling can indeed seem silly, but having used one or the other for some time and becoming proficient at it, and then hearing someone touting "your revered style is inferior and for amateurs, after all, real men use <the other style>" can lead to such hugely prolific wastes of time debating the various merits. And they're typically called religious arguments, as in the end, no one's opinion is really changed.

Indeed these are good times, with new models and machines coming out, some even offering multiple choices.

I say choose the one(s) you like and use them, leave the religious wars to those that like them.

But, I'll also add that having an infinite stack is FAR from understanding and using RPL. If you have an open mind and are curious, look into RPL, you may just like it. There are lots of books about it, but I recommend "HP 48 Insights Part I: Principles and Programming" availabe at https://literature.hpcalc.org/items/1571" by Bill Wickes.
--bob p

DM42: β00071 & 00282, DM41X: β00071 & 00656, DM10L: 071/100
Florian
Posts: 22
Joined: Mon Apr 26, 2021 1:19 pm
Location: Japan

Re: Inspiring the next generation of RPN users

Post by Florian »

Thanks a lot for explaining the origins of the 'repellent' word play. (I was aware of the latter but not of the former.) That makes things clearer. Still, even if intended in a joking mood, I feel that the phrase is partially used here as a catch-all in place of proper arguments that might be more productive.

I am aware that the N-level stack is but the tip of the RPL iceberg. Indeed, I am currently sweating my way through the 600 pages HP48 user's guide. This is to learn the basics of that machine before (perhaps) trying to tackle RPL. I agree, RPL seems more confusing to read. At least to the uninitiated (that is, my current me). The same probably also holds for keystroke programming. With the HP42S we are lucky to at least get the commands in human readable form. I imaging that those models that due to hardware limitations still required to be programmed and read in hex also had quite an interesting learning curve...
User avatar
dalremnei
Posts: 43
Joined: Thu Dec 10, 2020 11:32 am
Location: Scotland

Re: Inspiring the next generation of RPN users

Post by dalremnei »

Florian wrote:
Sun Nov 21, 2021 2:26 am
When I first obtained my DM42 it was, of course, a 4-level stack machine. This I had to accept as face value and I dealt with it. Then Thomas made his great move and gave us an N-level stack software and SwissMicros thankfully adapted it to their platform. For me personally that was a revelation. Suddenly the calculator felt natural to me. First from an aesthetic point of view. I only see numbers on the screen that I actually use. Additionally, having personally no need for the special nature of the T register, I was always rather annoyed by its behavior as it intended to "invade" my calculations with old numbers that I don't need any more to the point that I tried to do my calculation in three levels only. With the N-level stack I start with an empty stack and I know that at the end of the calculation I must be left with exactly one level left on the stack. If that is not the case I know that I made a mistake somewhere. That is reassuring. That is safety for me. That I required from a tool in a professional setting.

I use the calculator in my job as a physicist. In that it is a very rare case that I have an equation printed (or written) in front of me that I just need to solve. (This I can do much better, faster and safer with the PC in front of me.) Instead, the DM42 allows me to play with numbers, expressions, ways of approaching a problem. And now I can keep intermediate results visibly around to directly compare numbers. In that case I do not want to worry about running out of stack. As I don't yet know where my calculation is going to lead me I cannot pick the right position in the euqation so that I won't run out of stack. Therefore the N-level stack really opened the DM42 to me as a tool to explore physics in a very intuitive and sometimes more tangible way than my PC with Mathematica, Python and what not sometimes allows me to.
This has been my experience of NSTK as well. It has also allowed me to write a deadline tracker program that puts over 4 lines of information on the screen! Five lines was possible with careful use of the L register but going beyond was impossible. NSTK makes the DM42 a far more powerful calculator than it was before.

New users will prefer NSTK because it's just more intuitive. If I were marketing a hypothetical non-programmable RPN calculator to students, I'd rather say "just put in your operands and press the operator buttons" than "ENTER lifts the stack, oh but it's actually just 4 registers, and the T register gets copied to the Z register when it comes down, and if you run into really complex equations, you'll need to use the storage registers for intermediate results...". The dominance of mathprint calculators proves that users don't want to have to think about the implementation details of their calculators. They just want a tool that can be reliably used to solve problems. NSTK does that, within the confines of RPN.
SwissMicros DM42, DM16L, HP 12c Platinum, Prime G2 CASIO fx-9750gii, fx-991ex classwiz, fx-CG50, TEXET fx1500, TI nspire CX II-T
Post Reply