+- HP Forums (https://www.hpmuseum.org/forum)
+-- Forum: Not HP Calculators (/forum-7.html)
+--- Forum: Not quite HP Calculators - but related (/forum-8.html)
+--- Thread: newRPL - Ideas for "graphical" programming of math problems (/thread-16820.html)
newRPL - Ideas for "graphical" programming of math problems - Claudio L. - 04-27-2021 02:35 PM
Instead of a basic "FORMS" engine for newRPL, I'm thinking perhaps something with less programming would be more useful for the average user.
I'm welcoming ideas here on what this could look like.
The first thing that comes to mind is something like a MathCAD sheet, where a problem can be defined based on its variables and formulas in a visual way. Of course, more complex algorithms should be possible to code directly in RPL and simply use it in your formulae.
No, I do not want a spreadsheet, so that's out of the question.
Another possibility is something closer to LabView style of graphical programming, which allows for more complex algorithms to be coded in a graphical way, but it can get really tedious to code matrix indexing algorithms, so in the end it may not be that useful.
Has anyone seen of any other ways to represent math, physics or engineering problems without too much programming? I mean to draw new ideas to give form to something useful, if there's something cool out there that I've missed I want to know about it.
Originally, FORMS were simply a generic UI on top of the RPL environment. The main intent going in this different direction would be to make something more focused on solving problems (instead of just providing a UI to interface with the user).
RE: newRPL - Ideas for "graphical" programming of math problems - vanLudwig - 04-27-2021 04:11 PM
Have you seen the macro mode in Acron RPN Calculator? That's how I solved the same problem. It's not as powerful as a programmable calculator, but not so intimidating for average users.
http://www.acrongames.com/rpncalculator.html
RE: newRPL - Ideas for "graphical" programming of math problems - Claudio L. - 04-30-2021 09:16 PM
(04-27-2021 04:11 PM)vanLudwig Wrote: Have you seen the macro mode in Acron RPN Calculator? That's how I solved the same problem. It's not as powerful as a programmable calculator, but not so intimidating for average users.
http://www.acrongames.com/rpncalculator.html
It's an interesting take, but it seems to only allow you to replace constants and variables in a formula without re-typing it. I'm looking for something more along the lines of organizing the inputs to a problem, organizing the formulae and code needed to solve the problem, and organizing the solutions presented. A more 'human" way to code FORMS.
I think we can take a few hints from the HP Prime too, actually. For example the "Inference" app has 3 "Views". The Symbolic and Numeric view both make the user either select or provide some form of input variables (I consider the "method to use" also a variable of the problem here, for example). Then there's the "Calc" screen in the numeric View that presents some output and some other graphical output is presented in the Plot View.
I don't like the organization too much, to me it seems quite chaotic to find some results in one place, the graphic in a different place, some inputs in another place, etc.
In my opinion there should be 3 sections (views?): input data, solution, output. Solution would show you formulae and step by step resolution (for students, and also professionals that need to generate human readable output for others to review). In most cases you are interested in the final number, and you can skip to output.
But there's some concepts that are worth looking at on this "Inference" app of the Prime:
* One variable per row: We have vertical lists of items, each taking the whole width of the screen. It shows variable name on the left, a value on the right. For newRPL I'd like value on the center, units on the right for example. I like this organization, seems very clean.
* One "status line" at the bottom with a description of each line, displayed only when the user highlights this line.
* Some inputs are numeric variables, but some are selection lists, check boxes, etc. This is very flexible but it's really all the same: one variable taking a number with the index of the selection, or 1/0 if it's checked or not, just presented to the user with a different "look" in the UI.
* Each row/variable has a type (data validation), but it accepts flexible input, so if I want to enter a formula based on another variable I'm allowed to. For example in the same inference App I have to enter an integer for 'n' but I have a variable N with the value 10, I'm allowed to type N/2 as input. For newRPL, I'd like to accept the formula AND keep it as symbolic, only evaluating to a numeric type when it's time to recalculate. For example, if I type on the Prime N/2, the input variable n becomes 5 instantly. I'd like to keep the 'N/2' formula, so that if I change N=8, n would automatically recalculate as N/2=4.
Looking back to the Acron calculator macro mode you suggested, it seems a simplified "keystroke" version of what we see on the Prime, where you have a one-liner and change the inputs, while you see the outputs immediately displayed on the same row. This is quick to type and almost the same (as long as your whole problem is a single formula), but doesn't really allow you to package and publish that as an application.
RE: newRPL - Ideas for "graphical" programming of math problems - BruceH - 05-01-2021 12:25 PM
(04-27-2021 02:35 PM)Claudio L. Wrote: Has anyone seen of any other ways to represent math, physics or engineering problems without too much programming? I mean to draw new ideas to give form to something useful, if there's something cool out there that I've missed I want to know about it.
Perhaps the most enduring non-traditional way to program mathematics is A Programming Language (APL), invented by Ken Iverson in 1962, and still used today in a financial sector niche. There is a primer text here (PDF) and there is a modern variant called J which uses a traditional keyboard and character set.
For all the fuss about mini-Python on calculators, having a version of APL would be a cool touch - plenty of geeky kids would enjoy it - and it's ideally suited to a small screen because it's so compact. (In the PDF linked above there is an example 18-line BASIC program that is 5 characters in APL.)
RE: newRPL - Ideas for "graphical" programming of math problems - Columbus - 05-09-2021 08:16 AM
Hi, Claudio L.
If we are talking about visual programming methods, I would like to recommend you the DRAKON visual language. This language provides a uniform way to represent flowcharts of any complexity that are easy to read and understand. There is an article about this language on the English Wikipedia.
There are several IDEs for this language, the most famous of which are DRAKON Editor (offline version) and Drakon.Tech (web service).
(Sorry, I'm avoiding giving links to the ones I mentioned here, as I was unexpectedly banned for that a few days ago.)
RE: newRPL - Ideas for "graphical" programming of math problems - Columbus - 05-11-2021 05:40 PM
Hi, I'm Sergei, structural engineer from Russia.
I interesting in for graphing calculators. I have in my collection: TI-89T, HP Prime G2, HP 50g (ROM 2.15) and Casio Classpad 330 Plus.
To be honest, I'm novice in RPL (the most used calculator for me - TI-89T).
Now that concerns my post. As far as I understood, honoured Claudio L. set the question broader than something affecting only HP calculators on the RPL. He wrote:
Quote: Has anyone seen of any other ways to represent math, physics or engineering problems without too much programming? I mean to draw new ideas to give form to something useful, if there's something cool out there that I've missed I want to know about it.
That's why I turned his attention to a tool I know (on this subject). I think this corresponds to the subject of the discussion. I had no intention of spamming.
P.S. I didn't create a "new user ID" (by the way, what is that? login?). I just asked David Hicks to review my situation. And he made the decision to unban me (for which I thank him very much).
RE: newRPL - Ideas for "graphical" programming of math problems - Claudio L. - 05-13-2021 01:34 PM
(05-11-2021 05:40 PM)Columbus Wrote: Hi, I'm Sergei, structural engineer from Russia.
I interesting in for graphing calculators. I have in my collection: TI-89T, HP Prime G2, HP 50g (ROM 2.15) and Casio Classpad 330 Plus.
To be honest, I'm novice in RPL (the most used calculator for me - TI-89T).
Now that concerns my post. As far as I understood, honoured Claudio L. set the question broader than something affecting only HP calculators on the RPL. He wrote:
Quote: Has anyone seen of any other ways to represent math, physics or engineering problems without too much programming? I mean to draw new ideas to give form to something useful, if there's something cool out there that I've missed I want to know about it.
That's why I turned his attention to a tool I know (on this subject). I think this corresponds to the subject of the discussion. I had no intention of spamming.
P.S. I didn't create a "new user ID" (by the way, what is that? login?). I just asked David Hicks to review my situation. And he made the decision to unban me (for which I thank him very much).
Sergei, welcome to the forum, your post is appreciated here (moderators do their best to stop spammers, after you use the forums for a while you'll appreciate how clean they keep this forum, they do an amazing job even if sometimes it gets in the way of some users).
Back on topic, I looked at Drakon (I had never heard of it before). Seems like a generic flowchart for programming. It's very broad and general, maybe we can draw some ideas and make it more specific to math problems and engineering problems which is what I'm looking for: a way to represent your physics/engineering/math problems in a way to code less, rather than represent the code in a different way (which means you still have to code it).
Being a fellow structural engineer you know very well what I mean: you typically need to solve some specific, isolated problems repeatedly like a foundation pad with an eccentric load. I'm looking for a way to represent the problem itself (inputs, resolution, outputs) without the user having to code too much RPL. MathCAD does this very well, spreadsheets also work if you can get your head inside a cell.
RE: newRPL - Ideas for "graphical" programming of math problems - compsystems - 05-13-2021 01:46 PM
The development that SmathStudio is doing is very interesting, where the mathematical expressions are encoded in 2D in a similar way to MathCad
https://en.smath.com/forum/yaf_postst20283_Beta--SMath-Studio-0-99-7803--13-May-2021.aspx
Although for plain text we need a simple editor to code in RPL similar to HPuserEdit, since programming on the small screen of a calculator is too complicated.
https://www.hpcalc.org/details/6600
RE: newRPL - Ideas for "graphical" programming of math problems - Claudio L. - 05-14-2021 07:25 PM
(05-13-2021 01:46 PM)compsystems Wrote: The development that SmathStudio is doing is very interesting, where the mathematical expressions are encoded in 2D in a similar way to MathCad
https://en.smath.com/forum/yaf_postst20283_Beta--SMath-Studio-0-99-7803--13-May-2021.aspx
Although for plain text we need a simple editor to code in RPL similar to HPuserEdit, since programming on the small screen of a calculator is too complicated.
https://www.hpcalc.org/details/6600
Thank you, I was aware of SMathStudio as a MathCAD look-alike. I do think for problems the way MathCAD (and SMathStudio) works is one of the best as far as organizing the problem on a page.
I agree that a tiny screen is NOT a page so this style interface may not be completely applicable but perhaps a simplified/streamlined version of it. But I'm not giving up yet, there has to be a way to solve the problem without much coding.
I've also looked into children coding apps, the way they represent code with blocks so kids can easily join the blocks to make a robot move, for example. Some are simpler than others, and of course it's not really related to math but it's a fresh idea to represent blocks of either data input, calculations, actions, and even loops (blocked together). I think something like that could be doable (not for children, of course, just the blocks idea).
So you make a sequence of input blocks, calculations, and output blocks. Then you can group those in a bigger block, give it a name and you have a "function" block, that given the input as parameters, gives you the output back, and you can use those function blocks as black boxes in other more complex problems.
In the backend, the UI would of course convert blocks to RPL for execution.
RE: newRPL - Ideas for "graphical" programming of math problems - mfleming - 05-15-2021 12:07 AM
A great deal of research was done on Visual Programming in the 70's and 80's. There are collections of papers from conferences and journal special issues you could turn to for ideas. However I think the main conclusion was that visual representation worked best for specialized problem domains and grew cumbersome as the application of visualization was generalized. Not to sound discouraging but you may want to narrow the domain to matrix problems, physics, engineering, etc. I spent two years studying the field in the early 90's as a possible dissertation subject but gave up. If VP was easy then you could just copy someone else's solution
~Mark
RE: newRPL - Ideas for "graphical" programming of math problems - Edwin - 05-21-2021 03:00 PM
Personally, I also prefer it to be in the style of MathCad and Smath Studio. There is another similar program that is widely used in Engineering, the Engineering Equation Solver (EES)
https://en.wikipedia.org/wiki/Engineering_Equation_Solver
http://fchartsoftware.com/ees/
RE: newRPL - Ideas for "graphical" programming of math problems - Claudio L. - 05-21-2021 03:49 PM
(05-21-2021 03:00 PM)Edwin Wrote: Personally, I also prefer it to be in the style of MathCad and Smath Studio. There is another similar program that is widely used in Engineering, the Engineering Equation Solver (EES)
https://en.wikipedia.org/wiki/Engineering_Equation_Solver
http://fchartsoftware.com/ees/
Thanks!, never heard of EES. I looked at some of the videos and the UI seems pretty crude, not much to learn from it but it is indeed "visual" in the sense you type the equations but not code.
RE: newRPL - Ideas for "graphical" programming of math problems - CMarangon - 05-21-2021 06:22 PM
Hello!
I would suggest HTML 5, but this is not supported.
I think that HP calcs do not have support to PASCAL, BASIC or other language because they don't want to pay for copyrights, that is expansive. (this is my opinion only)
You do good programing making programs with many pictures, that make it looks like more interactive as possible.
By the way, Flash and Java was killed by newer browswers,
and it seems that similar things happens with HP calcs.
In the decade of 2010-2020 we had HP Prime that seems to run away from the traditional programming mode of the previous calculators, like HP48 series, including HP50.
By other side, there are people that think that HP Prime is the END GAME.
Who knows...
I hope to be wrong...
For this I cant give you a good answer about what kind of language to use.:-(
quote='Claudio L.' pid='148201' dateline='1621612172']
(05-21-2021 03:00 PM)Edwin Wrote: Personally, I also prefer it to be in the style of MathCad and Smath Studio. There is another similar program that is widely used in Engineering, the Engineering Equation Solver (EES)
https://en.wikipedia.org/wiki/Engineering_Equation_Solver
out of the traditional programming mode of the previous calculators
Thanks!, never heard of EES. I looked at some of the videos and the UI seems pretty crude, not much to learn from it but it is indeed "visual" in the sense you type the equations but not code.
[/quote]
RE: newRPL - Ideas for "graphical" programming of math problems - Sukiari - 05-24-2021 05:36 AM
What’s wrong with the 50g’s Equation Writer? I like it.
RE: newRPL - Ideas for "graphical" programming of math problems - Claudio L. - 05-26-2021 12:49 PM
(05-24-2021 05:36 AM)Sukiari Wrote: What’s wrong with the 50g’s Equation Writer? I like it.
Nothing wrong, but doesn't solve the type of problems we are looking to solve, it's only for typing and manipulating expressions.