Calling the HP50g numeric solver from within RPL?

[johan162]

As a part of a larger program I need to numerically solve a non-rational expression which can only be solved numerically so SOLVE wont work. I know I can setup and call MSLV and that works but it has a side-effect of displaying the convergence in the top row which is a bit of a nuisance since the solution is only an internal stepping stone and not of interest for the final calculations.

Is there some way to call the built-in single equation numeric solver (the one used in the forms based NUM.SLV)?

(.. and yes I know. I have already solved this with Maple or Matlab but for part practical and part historic reasons It would be really handy to have the resulting program either on my real HP50 or on my phones emulator for ease of access.)

?

[Juan14]

You need the command ROOT, it requires an algebraic expression or program in level 3: a variable in level 2: and a guess in level 1:. For example 'EXP(-X)-COS(X)' [SPACE] 'X' [SPACE]1[SPACE][ALPHA][ALPHA]ROOT [ENTER] returns 1.29269571937.

[Giuseppe Donnini]

See also this post for a similar use case.

[Joe Horn]

An alternative on the 50g is to use 30 MENU to launch the soft-key menu-based HP-48 style numeric interactive solver, which I prefer over the 50g full-screen inform-box style solver. You must store your equation in 'EQ' before executing 30 MENU. It's awesome. Try it.

Little-known hidden feature of the 50g solvers: when solving takes a long time, press any key (except ON), and you'll be able to watch the solver at work. This is especially useful if the solver is slowly diverging instead of slowly converging... it lets you abort the process early instead of wasting time waiting when there would be no solution forthcoming anyhow.

[johan162]

Thanks! Feel a bit ashamed ... Not sure how I managed to forget about the ROOT cmd.

... only mitigating circumstance is that my 50g been in my drawer for the last four years until we started to discuss old calcs at work last and I could show-off my old HP28S and wanted to show that the "modern" 50g could still be useful.

Too bad HP will never produce a "60G" or whatever a follow-up with full RPN would be called but I guess the business case is really gone (and the dev team ?) and you have to target the edu. market which I believe is the purpose with the Prime. I'm sure the prime is a marvelous machine so the fault lies with me but I couldn't get used to it and its now collecting dust in my drawer and I use my old 50g (or the emulator on a phone) for the few occasion where a hand calculator is simply the most efficient tool.

[Cristi Neagu]

(05-17-2020 09:44 PM)johan162 Wrote:  Too bad HP will never produce a "60G" or whatever a follow-up with full RPN would be called but I guess the business case is really gone

Sorry to hijack the thread a bit, but this is something that has been nagging me for a while now.

HP dropped the 50g. With it, they dropped the professional oriented RPN calculators and chose to focus on the educational oriented algebraic calculators. Nothing wrong with that, HP is trying to stay alive. But what that means is that they don't see a future in 50g style machines.

So... Can't we try and petition HP to consider releasing the source code for the 50g to the community under an open source licence? Then people could, for starters, fix the bugs that are still present. Then maybe it could be converted to run straight on the ARM processor, delivering a huge performance boost. That would be huge for the newRPL project, i would imagine.

[John Keith]

(05-18-2020 12:10 AM)Cristi Neagu Wrote:  HP dropped the 50g. With it, they dropped the professional oriented RPN calculators and chose to focus on the educational oriented algebraic calculators. Nothing wrong with that, HP is trying to stay alive. But what that means is that they don't see a future in 50g style machines.

So... Can't we try and petition HP to consider releasing the source code for the 50g to the community under an open source licence? Then people could, for starters, fix the bugs that are still present. Then maybe it could be converted to run straight on the ARM processor, delivering a huge performance boost. That would be huge for the newRPL project, i would imagine.

I largely agree with you and I would certainly like HP or someone (Swiss Micros?) to make a modern RPL calculator. Otherwise for HP to make a programmer friendly version of the Prime with full RPL.

However, NewRPL is a separate, platform-independent system written in C and not binary-compatible with Saturn code (i. e. No SysRPL).

[Cristi Neagu]

(05-18-2020 03:10 PM)John Keith Wrote:  However, NewRPL is a separate, platform-independent system written in C and not binary-compatible with Saturn code (i. e. No SysRPL).

Well, as you know, the HP 50g isn't running Saturn code directly either. It's running an emulator that is running the Saturn code. As far as i understand, the newRPL project is trying to achieve at least the same functionality as the 50g (with some extra things and a better interface) while running directly on the ARM processor. The obstacle is that they're doing it all from scratch. So what i meant by "would be huge for the newRPL project" is that they would at least have the source code to start from. I imagine it's much easier to create a Saturn code to ARM compiler than it is to rewrite the entire code by hand.

Of course, the whole thing is speculation on top of more speculation on my part.

[grsbanks]

(05-18-2020 12:10 AM)Cristi Neagu Wrote:  So... Can't we try and petition HP to consider releasing the source code for the 50g to the community under an open source licence?

The problem is that the biggest bug that I would love to see corrected personally (unreliable alarms and timekeeping) is not in HP code. It's in the emulator that, as far as I know, belongs to Kinpo.

[Cristi Neagu]

(05-18-2020 08:43 PM)grsbanks Wrote:  The problem is that the biggest bug that I would love to see corrected personally (unreliable alarms and timekeeping) is not in HP code. It's in the emulator that, as far as I know, belongs to Kinpo.

Well (and, again, this is all make-believe), if HP were to release the code without their emulator, presumably some bright spark out there could figure out a way to compile it for ARM directly, bypassing the Kinpo emulator.

[Marco Polo]

(05-17-2020 07:43 PM)Joe Horn Wrote:  An alternative on the 50g is to use 30 MENU to launch the soft-key menu-based HP-48 style numeric interactive solver, which I prefer over the 50g full-screen inform-box style solver. You must store your equation in 'EQ' before executing 30 MENU. It's awesome. Try it.

Little-known hidden feature of the 50g solvers: when solving takes a long time, press any key (except ON), and you'll be able to watch the solver at work. This is especially useful if the solver is slowly diverging instead of slowly converging... it lets you abort the process early instead of wasting time waiting when there would be no solution forthcoming anyhow.

The old good solver menu :-)
Much faster and effective that the Inform interface
Still using it, together with MSOLVR

AFAIK the feature of displaying the solver iteration was already on the 48sx. Not sure about 28s

[KeithB]

Why is there no museum entry for the 50G?

[Cristi Neagu]

(05-20-2020 08:26 PM)KeithB Wrote:  Why is there no museum entry for the 50G?

Too new, maybe... Not classic enough? Even though the 50g is discontinued and out of stock, i would imagine purists are in denial about the Prime and still consider the 50g to be the latest, current HP graphing calculator

[John Keith]

(05-20-2020 08:40 PM)Cristi Neagu Wrote:  Too new, maybe... Not classic enough? Even though the 50g is discontinued and out of stock, i would imagine purists are in denial about the Prime and still consider the 50g to be the latest, current HP graphing calculator

Actually, the purists are in denial about graphing calculators and consider the HP-41 to be the latest, greatest HP calculator.

[Steve Simpkin]

(05-20-2020 10:49 PM)John Keith Wrote:  

(05-20-2020 08:40 PM)Cristi Neagu Wrote:  Too new, maybe... Not classic enough? Even though the 50g is discontinued and out of stock, i would imagine purists are in denial about the Prime and still consider the 50g to be the latest, current HP graphing calculator

Actually, the purists are in denial about graphing calculators and consider the HP-41 to be the latest, greatest HP calculator.

HP-41? This new fangled "calculator" has an LCD display for crying out loud!
"Real" calculators use LED displays

[Massimo Gnerucci]

(05-20-2020 10:49 PM)John Keith Wrote:  

(05-20-2020 08:40 PM)Cristi Neagu Wrote:  Too new, maybe... Not classic enough? Even though the 50g is discontinued and out of stock, i would imagine purists are in denial about the Prime and still consider the 50g to be the latest, current HP graphing calculator

Actually, the purists are in denial about graphing calculators and consider the HP-41 to be the latest, greatest HP calculator.

Amen, bro.

[Thomas Radtke]

(05-18-2020 08:43 PM)grsbanks Wrote:  The problem is that the biggest bug that I would love to see corrected personally (unreliable alarms and timekeeping) is not in HP code. It's in the emulator that, as far as I know, belongs to Kinpo.

What is this bug? Does Emu48 have it, too? Else I imagine SM could create some hardware, no?

[grsbanks]

(05-21-2020 06:44 AM)Thomas Radtke Wrote:  What is this bug? Does Emu48 have it, too?

I don't know if Emu48 has it too but with the host hardware not being an actual calculator, I wouldn't expect it to work the same way.

The emulator on board the HP 50g screws with the clock as seen by the emulated hardware. In some cases the clock has been reported to jump backwards at times and if you set an alarm then turn the calc off, 99% of the time the alarm doesn't go off. Turn the calc back on and it "remembers" that there was an alarm that became due while it was off, and activates it.

[Thomas Radtke]

Ah, thanks! I've used the 18C appointment functionality back then. With modern technology, I forgot hat these calculators were kind of organizers, too.

[EngineerX]

(05-21-2020 06:44 AM)Thomas Radtke Wrote:  

(05-18-2020 08:43 PM)grsbanks Wrote:  The problem is that the biggest bug that I would love to see corrected personally (unreliable alarms and timekeeping) is not in HP code. It's in the emulator that, as far as I know, belongs to Kinpo.

What is this bug? Does Emu48 have it, too? Else I imagine SM could create some hardware, no?

Sorry this a a bit off-topic:

Does Emu48 use ROMs by HP?

Many thanks