Calculators you regret buying?

[Valentin Albillo]

.
Hi, Maximilian,

(12-10-2023 08:14 PM)Maximilian Hohmann Wrote:  my most memorable regret ever was an HP-48SX. I had read so much praise [...] But what a delusion! [...] and last but least, the abomination of RPL as a programming language. After trying it out for about two hours (one hour would already have been an inexcusable waste of lifetime…) I removed the batteries and never put them back in. This horrible experience dates back almost 20 years but I am still traumatized ;-)

+1

I fully agree, coincide and concur: RPL is an unholy abomination. Quoting myself from this 2018 thread:

"The problem with RePeLent is that it's almost impossible to not write unfathomable "garbage" even if you're trying hard not to be cryptic, and in any case the moment you resort to stack juggling and stack acrobatics for the sake of efficiency you'll immediately plunge head on on cryptography. [...] Stack juggling is virtually incompatible with "readability".

[... about Jupiter ACE's version of FORTH but fully applies to RPL as well ... ] [it] proved to be just too cryptic, it forced would-be programmers to take care of the low-level details (stacks, [...], etc) which is precisely the kind of thing a decent language should take care of so that the programmer can concentrate on the algorithms and high-level details of the task to solve, [...]


It's the programming language the one which should be working for me and releasing me of the low-level drudgery, not the other way around."

As for the OP's question I've never regretted buying a calculator because I first inform myself about its characteristics and look for sample code, timings, brochures, articles, opinions, etc. so I fully know what to expect and thus I've never been negatively surprised.

For instance, I never bought a TI calc several decades ago because I quickly realized they were physically inferior, feeble and unreliable, and I wouldn't touch AOS with a 10-ft pole.

Best regards.
V.

[carey]

(12-11-2023 11:05 PM)Valentin Albillo Wrote:  ...Stack juggling is virtually incompatible with "readability".

True, and all too common in single-file RPL programs. However, the best practice for writing RPL, demonstrated beautifully by D. R. Mackenroth (of HP) in his book HP48 Programming Examples is to create a separate directory for each program, consisting of a main program that calls several small subprograms (similar to sound Forth practice). Mackenroth gives a simple example of computing the hypotenuse of a triangle where the main program <<GetA GetB ComputeH Label>> is just a series of subprograms (verbs) that employ minimal stack juggling. Nice and readable!

(12-11-2023 11:05 PM)Valentin Albillo Wrote:  ...so that the programmer can concentrate on the algorithms and high-level details of the task to solve

It is interesting that this goal of focusing on high-level details can now sometimes be met best, not in a programming language, but in an unlikely place -- non-programmable scientific calculators like the Casio fx-991CW or TI-30X Plus. Consider Newton's method of root-solving. Sure, one can write a program for Newton's method, but these remarkable $20 non-programmables allow writing the recursive algorithm directly, e.g., Ans - f(Ans)/f'(x)|x=Ans and every time EXE is pressed (after first entering a starting value) the next iterate appears. Note that the high-level algorithm is written and displayed directly (no programming required). A second example is calculating the sum of square roots of the first n integers. Instead of writing a 7-step program, these calculators' Sigma operator allows the high-level algorithm to be written directly, replacing loops with summations.
n
Σ √i =
i=1
Just to be clear, I'm not suggesting the trivial option of just replacing programming with built-in functions. Instead, these scientific calculators and their operators allow writing high-level algorithms directly using math instead of computer science.

In keeping with the thread, my regret is buying a few extra programmable calculators when a non-programmable scientific calculator can sometimes be more fun!

[rprosperi]

Mackenroth's book is available here: https://literature.hpcalc.org/items/1575

[Eddie W. Shore]

The closest purchases to regrets that I have are:

The HP 49g+ (gold - early edition), TI-57 LCD, and TI-54, as they all have horrible keyboards.

I had a folding Casio programmable, I think it was the fx-795p, the battery just never lasted long.

There was one made by CVS long ago, it was a clone of the fx-300ms, but the keys were out of order and some functions were misspelled.

[Maximilian Hohmann]

Hello!

(12-12-2023 02:15 PM)Eddie W. Shore Wrote:  ... TI-57 LCD, and TI-54, as they all have horrible keyboards.

Oh yes, now that you remind me... For professional reasons my main collecting interest are navigation calculators, even more than calculators from HP, because they never made dedicated navigation models. Two very common models in the 1980ies were the Jeppesen AvStar and ProStar, both manufactured by Ti. The Avstar was based on the Ti-30LCD and the ProStar on the Ti-57LCD. I have several in my collection, but all of them are totally unuseable because of their bouncy keyboards. It is merely impossible to even add two three-digit numbers. Using them on board an aeroplane, where the result might really matter, would actually be foolish. Luckily I never bought one back in their day (german aviation authorities did not approve electronic calculators and made us use circular slide rules instead) because that would have resulted in instant regret! Similarly unuseable is the car computer from Ti, the "LogPit".

Regards
Max

[dlidstrom]

(12-10-2023 09:42 PM)polbit Wrote:  And yet still the best calculator HP ever made, and best small device programming language and environment ever devised by man

I am still in awe of how this system was conceived. How SysRPL and Saturn was implemented together is amazing, to me.

[Dave Britten]

(12-12-2023 05:38 AM)carey Wrote:  

(12-11-2023 11:05 PM)Valentin Albillo Wrote:  ...Stack juggling is virtually incompatible with "readability".

True, and all too common in single-file RPL programs. However, the best practice for writing RPL, demonstrated beautifully by D. R. Mackenroth (of HP) in his book HP48 Programming Examples is to create a separate directory for each program, consisting of a main program that calls several small subprograms (similar to sound Forth practice). Mackenroth gives a simple example of computing the hypotenuse of a triangle where the main program <<GetA GetB ComputeH Label>> is just a series of subprograms (verbs) that employ minimal stack juggling. Nice and readable!

This is the technique I try to employ when I need to write anything significant in RPL. Also use plenty of local variables. Whether or not that kind of design is appropriate for a handheld calculator is certainly up for debate, however.

[dlidstrom]

(12-13-2023 12:30 PM)Dave Britten Wrote:  This is the technique I try to employ when I need to write anything significant in RPL. Also use plenty of local variables. Whether or not that kind of design is appropriate for a handheld calculator is certainly up for debate, however.

What is the downside, then?

[polbit]

<<
"Wouldn't be a good month on HP Forums if it didn't involve some good old RPN vs RPL debate

Unfortunately we didn't have internet back in late 80s and nobody told me what an abomination RPL is, so I went ahead and purchased my 28C. Ever since then I have been an ignorant RPL user that has written many a program, obviously not realizing the error of my ways. If I only knew how wrong I was then..."
>>

I learned fairly quickly to heavily utilize the directory structure and reuse subroutines. With the dynamic stack I always found it very intuitive to program in RPL environment. Yes, readability can sometimes be an issue, but it just takes some practice to both design your programs to be clearer and to read them quickly. After a while it becomes second nature.

[Marco Polo]

(12-13-2023 01:21 PM)polbit Wrote:  <<
"Wouldn't be a good month on HP Forums if it didn't involve some good old RPN vs RPL debate

Unfortunately we didn't have internet back in late 80s and nobody told me what an abomination RPL is, so I went ahead and purchased my 28C. Ever since then I have been an ignorant RPL user that has written many a program, obviously not realizing the error of my ways. If I only knew how wrong I was then..."
>>

I learned fairly quickly to heavily utilize the directory structure and reuse subroutines. With the dynamic stack I always found it very intuitive to program in RPL environment. Yes, readability can sometimes be an issue, but it just takes some practice to both design your programs to be clearer and to read them quickly. After a while it becomes second nature.

It is quite interested to observe how much RPL is not liked by RPN user, but the contrary is generally not true :-)

[Eddie W. Shore]

I just remembered: another calculator that I regret buying is a solar scientific calculator in the form of a rule sold by Staples. I don't think Staples sells it anymore.

Pressing [2nd/Shift] [ ON ] is supposed to turn on the stats mode and it doesn't. Further more, four of the keys were in the form of arrow keys.

https://edspi31415.blogspot.com/2015/10/...tific.html

[Maximilian Hohmann]

Hello!

(12-13-2023 01:57 PM)Marco Polo Wrote:  It is quite interested to observe how much RPL is not liked by RPN user, but the contrary is generally not true :-)

I think the RPL calculators would have had more success if "classic keystroke programming" would have been offered as an alternative to RPL (like the switching between RPN and algebraic mode on later models).
All the users of previous machines (from HP65 to 41/42) would have instantly felt at home on these calcuiators and enjoyed the extra features like unlimited stack, graphics and CAS. But the prospect of having to work one's way through 1000 pages of manual before being able to program the calculator turned away a lot of people, many of them for good.

Regards
Max

[polbit]

(12-13-2023 02:41 PM)Maximilian Hohmann Wrote:  Hello!

(12-13-2023 01:57 PM)Marco Polo Wrote:  It is quite interested to observe how much RPL is not liked by RPN user, but the contrary is generally not true :-)

I think the RPL calculators would have had more success if "classic keystroke programming" would have been offered as an alternative to RPL (like the switching between RPN and algebraic mode on later models).
All the users of previous machines (from HP65 to 41/42) would have instantly felt at home on these calcuiators and enjoyed the extra features like unlimited stack, graphics and CAS. But the prospect of having to work one's way through 1000 pages of manual before being able to program the calculator turned away a lot of people, many of them for good.

Regards
Max

It it certainly up to you to "turn away, many for good" because RPL is sufficiently different from RPN that it requires some time commitment to learn and become proficient in. That doesn't however make RPL an "abomination" because you spent whole of two hours playing with an old 48SX and didn't grasp it/like it.

As far as success in the marketplace, RPL calculators succeeded no matter how you look at it. Spanning 28 years from 28C in 1987 to 50G in 2015, and having many, many dedicated resources like its' own newsgroup comp.sys.hp48, to hundreds of books, user groups, etc.

[Maximilian Hohmann]

Hello!

(12-13-2023 03:20 PM)polbit Wrote:  ...That doesn't however make RPL an "abomination" because you spent whole of two hours playing with an old 48SX and didn't grasp it/like it.

I understood this thread as an opportunity to share the (very personal!) feeling of "regret" after purchasing a calculator. At the time I bought the HP-48SX I had earned my living using and programming calculators and computers for several decades. Using programming languages and paradigms including calcultor keystroke programming in AOS and RPN, Basic, FORTRAN, C, C++, Pascal Mathematica and Java. For example, coming from C++, I did not even need a book to be able to write programs in Java... So when I got the HP-48 I thought I could use it and program just like any other HP I had come across. But no, they tried to invent a different kind of wheel. HP, no longer existing as a manufacturer of calculators, may have been able to sell their special kind of wheel for 30 years, but everybody else (Sharp, Casio, Ti, ...) continued to make round wheels and sold and still sell several orders of magnitude more of their product.

Regards
Max

[Marco Polo]

(12-13-2023 02:41 PM)Maximilian Hohmann Wrote:  Hello!

(12-13-2023 01:57 PM)Marco Polo Wrote:  It is quite interested to observe how much RPL is not liked by RPN user, but the contrary is generally not true :-)

I think the RPL calculators would have had more success if "classic keystroke programming" would have been offered as an alternative to RPL (like the switching between RPN and algebraic mode on later models).
All the users of previous machines (from HP65 to 41/42) would have instantly felt at home on these calcuiators and enjoyed the extra features like unlimited stack, graphics and CAS. But the prospect of having to work one's way through 1000 pages of manual before being able to program the calculator turned away a lot of people, many of them for good.

Regards
Max

Caveat: I use both RPL and RPN calculators.
I bought my first HP in 1989, when i was a chemical engineering student, a 28s.
At that time I did not know the difference between RPN and RPL.
As a matter fact it was the first time I met something different than algebraic logic.
After a short training with postfix notation I started exploring the calculator and fell in love with it.
The directories allowed me to keep things in order, the programming capabilities where amazing to me (I was used to Turbo Basic, Fortran77 and Turbo Pascal): the calculator was able to perform many calculation task easier than a PC with a Fortran compiler. Basically, I graduated almost not using a PC.
The comparison with a 42s of one of my mates, had a clear winner: the 28s was much more flexible and capable to handle complex calculations with ease, while the 42s required more effort to develop complex programs.
After one year I switched to the 48sx, my mate sold the 42s and bought my 28s.

And what about RPN? I use and program a 35s (and Plus42) daily, but find more intuitive to program the 48/50.
With such calculator I can do something similar to keystroke programming.
I just type << and start pushing keys.
The use of global or local variables greatly reduces the need of stackrobatics.
I admit that RPL legibility may be difficult, but IMHO is just due to the "presentation" of the program as a long string instead of one instruction per row as in RPN.

All in all, both systems have pros and cons. It may be a matter of personal tastes and/or of task to perform.
In my case, they can cohexist.

[polbit]

(12-13-2023 06:34 PM)Maximilian Hohmann Wrote:  Hello!

(12-13-2023 03:20 PM)polbit Wrote:  ...That doesn't however make RPL an "abomination" because you spent whole of two hours playing with an old 48SX and didn't grasp it/like it.

I understood this thread as an opportunity to share the (very personal!) feeling of "regret" after purchasing a calculator. At the time I bought the HP-48SX I had earned my living using and programming calculators and computers for several decades. Using programming languages and paradigms including calcultor keystroke programming in AOS and RPN, Basic, FORTRAN, C, C++, Pascal Mathematica and Java. For example, coming from C++, I did not even need a book to be able to write programs in Java... So when I got the HP-48 I thought I could use it and program just like any other HP I had come across. But no, they tried to invent a different kind of wheel. HP, no longer existing as a manufacturer of calculators, may have been able to sell their special kind of wheel for 30 years, but everybody else (Sharp, Casio, Ti, ...) continued to make round wheels and sold and still sell several orders of magnitude more of their product.

Regards
Max

I think there is a much better comparison with programming languages, as I look at RPN as good old K&R C with some inline assembly, whereas RPL would be more akin to C++. The transition from structured to object-oriented programming was quite dramatic for a lot of people, yet necessary for progress as complexity increased. HP did improve the wheel with RPL, and did so very successfully in the professional space.

Anyway, you are correct in that this thread started as sharing stories of regret in calculators. I shall refrain from turning it fully into another RPL debate. Cheers!

[polbit]

(12-13-2023 06:37 PM)Marco Polo Wrote:  All in all, both systems have pros and cons. It may be a matter of personal tastes and/or of task to perform.
In my case, they can cohexist.

HP 48SX with upgraded screen, and with the 41CV Emulation ROM and overlay. Problem solved

[aurelio]

I regret to have bought my first HP calculator on ebay, starting to play as a "little collector", that it's quite different to be simply an utilizer of them,

for the same reason I could say that I've regret so to have no bought a lot of calculator, when I had the chance to, expecially the ones lost in a last minute competition auction

I reply only here to both the reladed topics, just to save money for the next purchase

[Marco Polo]

(12-13-2023 07:26 PM)polbit Wrote:  

(12-13-2023 06:37 PM)Marco Polo Wrote:  All in all, both systems have pros and cons. It may be a matter of personal tastes and/or of task to perform.
In my case, they can cohexist.

HP 48SX with upgraded screen, and with the 41CV Emulation ROM and overlay. Problem solved

In my case: 50g at office, 35s in backpack (aka "mobile office"), Emu48 (with my 50g clone) and Plus42 on smartphone and laptop

[Marco Polo]

(12-13-2023 07:05 PM)polbit Wrote:  Anyway, you are correct in that this thread started as sharing stories of regret in calculators. I shall refrain from turning it fully into another RPL debate. Cheers!

Just to get back in topic, I bought in 2004 a 33s.
Never got used to the chevron keyboard, cluttered print on bezel and unreadable (to me) color scheme.
Sold to buy the 35s in 2007.

One more, a 49g+ bought in 2005: one key broke, calculator replaced, broken again in 2007 and replaced by HP with my actual 50g after many discussion (apparently the fact the calculator was for professional use convinced HP for a replacement with upgrade....)