Request for suggestions: algebraic equations in Free42

[Marco Polo]

(07-08-2020 09:16 PM)Thomas Okken Wrote:  like a big screen, big stack, units, and algebraic equations.

Disclaimer: no polemic intent, just my two cents :-)

Frankly speaking, a Free42 with such improvements seems to me very near to a 48s. Just add directories and you will get a "RPN 48s"....
I really do like the concept, especially if loaded on DM42 hardware (and probably i would buy one), but is it really necessary to reinvent the wheel?

[Vincent Weber]

(01-12-2021 02:55 PM)Marco Polo Wrote:  

(07-08-2020 09:16 PM)Thomas Okken Wrote:  like a big screen, big stack, units, and algebraic equations.

Disclaimer: no polemic intent, just my two cents :-)

Frankly speaking, a Free42 with such improvements seems to me very near to a 48s. Just add directories and you will get a "RPN 48s"....
I really do like the concept, especially if loaded on DM42 hardware (and probably i would buy one), but is it really necessary to reinvent the wheel?

Well, not quite. The 42S is programmed with RPN, much simpler than RPL, and with a 4-level stack (with automatic replication of the T-register) rather than an "infinite" one.
I believe a lot of people of this forum will find this reason alone good enough to stick with free42, no matter the enhancements.

[Thomas Okken]

The next version will have an unlimited stack. (As an option, the default will still be the four-level stack! That will always be necessary because lots of legacy programs won't work with the unlimited stack.)

The idea is to combine the ease of RPN keystroke programming with the power of RPL, or at least a few key parts of it.

Whether this will run on the DM42 remains to be seen, I think RAM is a bit tight even now, so it might not be practical.

[Marco Polo]

(01-12-2021 05:10 PM)Thomas Okken Wrote:  The next version will have an unlimited stack. (As an option, the default will still be the four-level stack! That will always be necessary because lots of legacy programs won't work with the unlimited stack.)

The idea is to combine the ease of RPN keystroke programming with the power of RPL, or at least a few key parts of it.

Whether this will run on the DM42 remains to be seen, I think RAM is a bit tight even now, so it might not be practical.

It would be a very interesting calculator, an ideal merge of two universes.
But, for some strange universal law, the ram is always almost full..... :-D

Is a beta version already available for testing?

[Thomas Okken]

(01-12-2021 06:19 PM)Marco Polo Wrote:  ram is always almost full..... :-D

Not when you run it on a computer or a smartphone.

(01-12-2021 06:19 PM)Marco Polo Wrote:  Is a beta version already available for testing?

Not yet, there is still a lot left to do. But there will be a version with just the big stack soon, I hope in a week or two.

[rprosperi]

(01-12-2021 05:10 PM)Thomas Okken Wrote:  The next version will have an unlimited stack. (As an option, the default will still be the four-level stack! That will always be necessary because lots of legacy programs won't work with the unlimited stack.)

Will the new Free42 version with this and other new features have build options to include them or not, so if, for example, the DM42 does not have sufficient RAM to host the unlimited stack feature, can that option simply not be configured in the build (and then obviously not exposed in the program), to allow the core Free42 with other enhancements continue to be delivered in DM42 f/w as new versions/features emerge?

[Thomas Okken]

The big stack feature will only use a few bytes of RAM when not in use, so that shouldn't be an issue, I imagine. It will make the code larger, so it will use more flash, and no, it won't be a compile-time option.

[richmit]

This isn't a suggestion. More of a story.

I have a mathematical software library that can take a function, and do things like find roots, approximate numerical derivatives, integrate, find extrema, etc... The whole thing is in LISP, and I wanted to be able to feed it functions defined as formulas using traditional infix notation. What did I do?

I wrote another function that would take a string containing an equation, and then burp out a LISP function that implements the equation. LISP is funny like that. It is super easy for a program to program itself.

Seems like that would be pretty easy for Free42. That is to say use Dijkstra's algorithm to convert the equation to post-fix (RPN), add a label at the top & some MVARs. Done.