Page 55 of 96

Re: 43S Alternative key layout --> WP43C

Posted: Tue Oct 01, 2019 7:44 am
by Dani R.
H2X wrote:
Mon Sep 30, 2019 8:52 pm
...

Maybe an undo function might make as much sense as a corrective operation as a E- key? Assuming you realise that you made a mistake right away, it would be easier to press one key to undo, than to repeat the value and then press E-.
Σ-, I think, is not just used as undo. In the history of statistics there are always people who want to prove their prejudice. So if the expected result doesn't fit yet, you can remove the data point you suspect to spoil the nice result with Σ-.

Re: 43S Alternative key layout --> WP43C

Posted: Tue Oct 01, 2019 7:55 am
by inautilus
Dani R. wrote:
Tue Oct 01, 2019 7:44 am
Σ-, I think, is not just used as undo. In the history of statistics there are always people who want to prove their prejudice. So if the expected result doesn't fit yet, you can remove the data point you suspect to spoil the nice result with Σ-.
+1

Any fool can manufacture a preordained result ... :D

Re: 43S Alternative key layout --> WP43C

Posted: Tue Oct 01, 2019 11:00 am
by H2X
inautilus wrote:
Tue Oct 01, 2019 7:55 am
Dani R. wrote:
Tue Oct 01, 2019 7:44 am
Σ-, I think, is not just used as undo. In the history of statistics there are always people who want to prove their prejudice. So if the expected result doesn't fit yet, you can remove the data point you suspect to spoil the nice result with Σ-.
+1

Any fool can manufacture a preordained result ... :D
Excellent! You just smashed the case for getting rid of E- altogether! :-)

On the WP43S/C, that is. Other (statistical) calcs can have their cake and eat it...

Re: 43S Alternative key layout --> WP43C

Posted: Tue Oct 01, 2019 12:06 pm
by Jaymos
H2X wrote:
Tue Oct 01, 2019 11:00 am
inautilus wrote:
Tue Oct 01, 2019 7:55 am
Dani R. wrote:
Tue Oct 01, 2019 7:44 am
Σ-, I think, is not just used as undo. In the history of statistics there are always people who want to prove their prejudice. So if the expected result doesn't fit yet, you can remove the data point you suspect to spoil the nice result with Σ-.
+1

Any fool can manufacture a preordained result ... :D
Excellent! You just smashed the case for getting rid of E- altogether! :-)

On the WP43S/C, that is. Other (statistical) calcs can have their cake and eat it...
I like Dani’s point, but in reverse and not as he intended it. In a real world with real instruments and real humans, I have found that I have to sometimes discard bad samples. A statistical machine needs the possibility to do that. The user decides (on prejudice) on whether data must come out of the set, and not the machine designer. I still want E-!

Re: 43S Alternative key layout --> WP43C

Posted: Tue Oct 01, 2019 12:20 pm
by Dani R.
Translations are always points of misunderstanding.

As long as we support linear regressions on the calculator, we also need the function Σ-.

In general, I think that we should leave functions in the menus if they don't interfere too much, even if there is a direct button. Thanks to the User Mode, almost every key can be assigned differently. If you have now assigned a key differently, it is quite practical to find the original function also in the menu, which fits to the function. Otherwise you would have to deactivate the User Mode, press the key, and then activate User Mode again. Or XEQ "Alpha". Or CATALOG -> FNCS.

Re: 43S Alternative key layout --> WP43C

Posted: Tue Oct 01, 2019 4:20 pm
by inautilus
Image

Re: 43S Alternative key layout --> WP43C

Posted: Tue Oct 01, 2019 8:18 pm
by Guenter
Jaymos wrote:
Tue Oct 01, 2019 12:06 pm
H2X wrote:
Tue Oct 01, 2019 11:00 am
inautilus wrote:
Tue Oct 01, 2019 7:55 am


+1

Any fool can manufacture a preordained result ... :D
Excellent! You just smashed the case for getting rid of E- altogether! :-)

On the WP43S/C, that is. Other (statistical) calcs can have their cake and eat it...
I like Dani’s point, but in reverse and not as he intended it. In a real world with real instruments and real humans, I have found that I have to sometimes discard bad samples. A statistical machine needs the possibility to do that. The user decides (on prejudice) on whether data must come out of the set, and not the machine designer. I still want E-!
And it is necessary if you just made an error by inputting a wrong data point. Simply take this data point again key ∑- and you're done. Otherwise you'd have to start over and key in all the data points again.

And when you are examining different scenarios it would be very handy to have that ∑- directly available.

Günter

Re: 43S Alternative key layout --> WP43C

Posted: Tue Oct 01, 2019 8:21 pm
by Jaymos
Guenter wrote:
Tue Oct 01, 2019 8:18 pm

And when you are examining different scenarios it would be very handy to have that ∑- directly available.
True, good point.

Re: 43S Alternative key layout --> WP43C

Posted: Tue Oct 01, 2019 11:32 pm
by H2X
Guys, I wasn't saying that there shouldn't be a corrective action, I was just suggesting that E- might have flaws and there might be better alternatives. I did mention undo, and others mentioned or hinted list / matrix editing.

Who needs to be able to remove values which have not been entered in the first place, or the bother of typing exact values again with the possibility of typing them wrong?

Re: 43S Alternative key layout --> WP43C

Posted: Wed Oct 02, 2019 12:40 am
by Jaymos
Hi all

I updated the 43C source to the main project's latest, thanks to Martin. I tested the CEIL, FLOOR, CUBEROOT, IP and FP changes and found no issues.

Furthermore, re 43C,

I rearranged the STAT menu as I thought best without upsetting the structure too much. I think it is nice now. See picture below.

Regarding the COMPLEX / CC discussion, I added two default USER mode assigns (for test purposes only). The ASN command is not working, I hard coded these defaults, and plan to take it out when the main project ASN is working:
  • COMPLEX is on the normal position, now HP42S compliant, i.e. it will convert x+iy to x and y on the stack and vice versa.
  • in USER mode ( f[2] ), E+ changes to CC. CC is the WP43S CC, which does both interactive and x+iy conversion. As previously.
  • in USER mode ( f[2] ), f[E+] changes to CC1. CC1 is interactive entry ONLY. I don't particularly see a need for this one. I would like comments whether I should even keep CC1.
You can toggle USER mode on and off by f[2]. (If there is interest, I could hard code some more temporary USER mode re-allocations for testing. Just let me know).

When Dani compiles the Windows Emulator and sends it to me tomorrow, I will load it on my http server and make it available. Chances are 96.34 % that the emulator may be available tomorrow morning ;-)

Edit: PS. The backup.bin file MUST be deleted prior to running.

Code is pushed up to the import5 branch.
DM42 image: (I did not yet test it, sorry, just compiled, no time): http://cocoon-creations.com/download/43C_13L2.pgm
Picture:
Image