43S Alternative key layout --> WP43C

[Jaymos]

Has anyone mentioned pressing and holding a key for its second function? I use this ask the time on RealCalc (RPN mode, naturally) and I find it a real time saver. It'd love a hardware RPN calculator that does the same thing.

Hi,

Yes, that is the so called H2X solution, i.e. holding in, until it changes on screen to second function f, then third function g.

I will make a test case on the emu and real calc to do that.

But first some other important work to be finished before I can do that ... a week or so should be ok.

J

[Jaymos]

This Christmas release version of the WP43C demo work in progress is far from a polished product. This demo comprises:

1. The new double precision only WP43S calculator engine.

2. The new fixed input types.

3. The new function key softmenu underlining

4. The function keys now can be operated by the usual f and g preceding Fx, but also by holding in the function key Fx until the correct option appears.

5. The shift modes f, and g can now also be advanced in the cycle from f to g to HOME using long presses. No mode change is required. Short press, double press and long press modes simultaneously work.

6. This release provides visual feedback to the lively discussion we had on this forum regarding long presses, double presses and the like. The longpress method of selecting the FN keys works reasonably well and can be demonstrated by this version. The double press function does not work well. Double press can be selected in the MODE softmenu, but it seems that the DM42 cycle time could be too slow for proper double click recognition. Either way the double press does not work reliably and instead of selecting g[Fx] which is intended by the double press, often Fx is selected.

Most certainly the multitude of changes in this version brought with some unwanted bugs, but I am intent on finding those. In fact there is a list of bugs I am still hunting down, but disregarding this, the version is good to demo most things. Apart from a weird double press to OK an option, Short Integer as input type still does not work, Comments, bug reports and suggestions welcome as always;-)


I updated the differences document:
http://cocoon-creations.com/download/W ... s_v045.pdf


Dani updated the Windows Emulator:

http://cocoon-creations.com/download/WP ... -12-24.zip
Remember to delete the backup.bin file prior to running the new WP43C emulator.


DM42 binary

http://cocoon-creations.com/download/43C_20L2.pgm
Remember to delete the settings using CLEAR/RESET when running the DM42 version.7


Video showing the new features.

This is a silent video as I was in a very noisy environment making the video. Some other day I will talk you through another video with sound, but for now, the video suffices to demo the underlining effect indicating f and g shifts at the FN keys and the input types .

https://youtu.be/jyHXniR-f_c


Geseënde Kersfees
Regards
Jaco

[Jaymos]

Hi all,

Here is an interesting little module I added to the calculator.

It is in anticipation for 43S SOLVER, with formulas to be added later. The intention is to plot a formula which is entered into the SOLVER, or for that matter entered as a RPN program. The structure does not matter, because all is needed is an RPN program with X input and X output.

Either way, I did this example using two simultaneous simulated (hard coded) RPN formulas to test the graphics code:
The RPN formula structure is clear from the C code below, and both Lbl_A and Lbl_B formulae are simultaneously plotted.

The emulator takes about half a minute to run the graph, and the DM42 takes a lot longer, maybe 3 times longer. I'll do a time check if there is interest.

Code: Select all

void Fn_Lbl_A(void) {
    fnAngularMode(AM_RADIAN);                           //Does not belong here -- it is repeated. It is convenient.
    copySourceRegisterToDestRegister(REGISTER_X, 99);   // STO

// (sin(x)/x + sin(10x)/5) / 2 + 2/x
 
    fnSin(0);         // SIN
    fnRCL(99);        // RCL 99
    fnDivide(0);      // /
 
    fnRCL(99);        // RCL 99
    fnStrInput2("10");// 10
    fnMultiply(0);    // *
    fnSin(0);         // SIN
    fnStrInput2("5"); // 5
    fnDivide(0);      // /
 
    fnAdd(0);         // +    

    fnStrInput2("2");// 2
    fnDivide(0);     // /

    fnStrInput2("2");// 2
    fnRCL(99);       // RCL 99
    fnDivide(0);     // /
    fnAdd(0);        // +    
}


void Fn_Lbl_B(void) {
	fnAngularMode(AM_RADIAN);   //Does not belong here -- it is repeated.

    fnStore(99);
    fnSin(0);                // SIN    
    fnRCL(99);
    fnSquare(0);           // square
    fnSin(0);                // SIN
    fnAdd(0);               // +    
}

Clipboard25.png (3.51 KiB) Viewed 267 times

[Jaymos]

I played some more with the code and made it a bit neater. I added a basic interface to adjust the axis.

The ticks are auto generated if TICK DX and DY are 0 otherwise you override the auto setting by entering values from the keys.

This module is now parked until we have a way to program a formula, either by SOLVER or by RPN sequence. I imagine I would put a text input corresponding to the RPN label for graph functions 1 and 2, to be plotted on the same set of axis.

Clipboard27.png (3.75 KiB) Viewed 242 times

The scaling is done by the values in the screen. The values are in the sequence xmin, xmax ... as in the screen text which comes onto the stack if you press LIMITS.

Clipboard26.png (2.31 KiB) Viewed 242 times

The softmenu buttons are used to enter the limits on the axis, i.e one would do 0.5 [CHS] [Ymin] to change the bottom on the y axis from the current -1.0 to -0.5, and 40 [Xmax] to make the positive x axis cut off at 40. The second graph is for these values:

Clipboard28.png (2.42 KiB) Viewed 242 times

[H2X]

Great work!

[Jaymos]

Thanks H2X

If all goes well I will release a simulator file and DM42 firmware version tomorrow. There was lot of config menu changes (driven by 43S and that we don‘t have an additional key space available for their new CFG menu), a small key label change (after I actually fixed the BASE menu), and Dani added binary checkboxes, and I added the graph menu.

I can report that I really like the feel of the long press FN keys with the underlining showing the actual function key. It works well and creates an impression of feeling in control with the visual feedback. I also like the fact that the FN keys now also time out to NOP which also strangely creates good feedback/control.

I need to take another stab at the FN double click which accelerates (and short cuts) the long press period to get to g(FN). That currently does not work and I don’t have time for attempting that again now.

(The complication with that is that as soon as FN key is pressed, the processor must register and start timing out primary FN key activation with a non-noticable delay of no longer than say 0.1 sec, and simultaneously check whether a key depress-and-repress cycle happens within that time period (or shortly thereafter), and it must still cycle through normal processing loop (which eats inconsistent time) to maintain the order and to also display the command name top left. Loop processing must be maintained without the availability of interrupts. A tall order actually, and that leaves out the simulator which IS interrupt driven and must also work in the same code, and GTK has it’s own fair share of quirks . But let me not confuse my problems with yours )

Great work!

[Jaymos]

The promised update.

This version contains configuration menu updates, SHOW test, REGS.V update, a key update and the graph module.

The graph module test (see g[9] Up), displays the standard demo SIN(x)/x graph and I will continue with that development once we get a way to include a labelled RPN sequence, which is (afaik) not in process at the moment.

A small keyboard change happened, i.e. I decided that the d.ms label on the top row of the keyboard is not relevant there and I moved it to be with the other Angle Display Mode selections in the MODE softmenu, i.e. DEG/RAD, etc. That meant that on the emulator layout (L1A) I could move SAVE up from g[1] to g[sqrt] and use g[1] for the BASE menu.

On the DM42 Layout 2 (for the DM42), the BASE menu and SAVE commands have previously been allocated, therefore the g[sqrt] position is now open and available for a sensible change. I was contemplating putting factorial (x!) back on the L2 keyboard in this position. Either way, the new label should preferably be a math function not a command, and definitely not a MENU call, hence the idea of x!. Any ideas? This change has not been done.

Furthermore, the WP43S SHOW command is in pre-alpha state as per the development of the WP43S, which is always integrated into my WP43C releases. You can check that in the simulator.

I updated the differences document, added a few pages and changed a few pages. There is now a table detailing the keyboard differences between WP43C L1A, WP43C L2 and WP43S. There is also a graphical representation and discussion of the individual configuration options (where different from 43S).

The Windows binary executable, together with DM42 firmware version 43C21L2.pgm file are in the ZIP FILE.

This week I will still change the configuration menus (MODE, DISP and CFG) a bit, but the structure as in the simulator and document is pretty close to useful. I do foresee a little shuffling still.

[Dani R.]

On the DM42 Layout 2 (for the DM42), the BASE menu and SAVE commands have previously been allocated, therefore the g[sqrt] position is now open and available for a sensible change. I was contemplating putting factorial (x!) back on the L2 keyboard in this position. Either way, the new label should preferably be a math function not a command, and definitely not a MENU call, hence the idea of x!. Any ideas? This change has not been done.

nth root of a number x, e.g xth root of y.
3th root of -8 will then give -2.

[Dani R.]

The item g[CLK] seems to have little to do with "clock". The main menu has date and time related items.

Would it be better to name it "TIM" or "TIME"?


Ideas?

Layout 1A: You could now exchange BASE (g[1]) and CLK (g[4]). Then BASE, BITS and INTS are on the same row. I haven't found a better abbreviation for CLK yet.

Soft menu EXP: I think SQRT belongs on f[F3], even if it seems superfluous. But maybe someone comes up with the idea to assign SQRT differently in USER mode and is then happy to find SQRT in the soft menu EXP. In the catalog FCNS it is a bit tedious to find SQRT.
In row G you should sort "sinh..artanh" like in the soft menu TRIG.

[Jaymos]

Hi

Since the previous release of Layout 1A, we have not changed it.
Last week, we added the x-th root of y instead of d.ms, and the replacement of SAVE with BASE. Therefore we should rename after these changes today.

So, for the new revised Layout (1B):

That is a good idea, swapping those BASE & CLK. Also, I can combine this with another idea that I had this weekend, and that is to combine CONV and UNIT. I feel that units are conversions, and we can combine it without the loss of understanding.

1. BASE, BITS and INTS are next to one another
2. UNIT is inside CONV
3. aSTR is with Alpha
4. The MULpi convert buttons are in the same sequence as per the DISP setting page of DEG/RAD...
5. The UNIT menu must re-arrange to have the 6 most common imperial/metric conversion on the primaries. I guess that could be the ones I indicated?

Thoughts?

Clipboard35.png (59.43 KiB) Viewed 49 times

EXP menu hyperbolics: I agree. Will re-arrange hyperbolics to be the same sequence as the hard keys.
EXP menu SQRT: I fully agree it completes that root sequence. I will change it.

Regards
J