CW 32B work thread

[Joseph_21sv]

Here are the preliminary specifications of the CW-32B
Display size: 7x83 dot matrix+15+3 seven segment digits
Display color: Black/Red/Green
Numerical precision: 15 digits
Entry method: RPN, but dot matrix can be set to display textbook notation
Memory: 74 numbers [A…Z, a…z, 0…9, .0….9, theta, (I)]+144 numbers addressed by (I)
Memory arithmetic: + - * / ^
Mode settings: non-volatile
Basic arithmetic + - * / ^ with shortcuts to x^2 sqrt(x) x! % %change %total
Special functions:
Redefinable keyboard: yes
Shift keys: yes
Alphabetic characters: accessible by ALPHA key
Exponentials, Logarithms: base e, 10, y
Trigonometrics: circular, hyperbolic, shortcut for pi tau=2*pi eta=pi/2 sigma=pi/4, degree and radian measure
Statistics:
Types: x, x-weight, x-y, x-y-weight, x-x1-y
Storage: 96 data points
Summation: x, x1, y, x*x1, x*y, x1*y, x*x1*y, x^2, x1^2, y^2
Measures of central tendency: Mean x, x weighted, x1, y, y weighted; population/sample standard deviation x, x weighted, x1, y, y weighted; population/sample standard error x, x weighted, x1, y, y weighted; covariance; Minimum, Q1, Median, Q3, Maximum x, x1, y
Regression models: O(f(x)^4|f(x) in {x^a, 1/O(x^4), e^ax, ln(O(x)), sin(O(x)), sec(O(x)), sinh(O(x)), sech(O(x))})
Probability distributions: same as WP-34S
(Pseudo-)Random Number Generator
Date entry: month/day/year, day/month/year
Date calculations: 30/360, actual/actual
Menus/Prompts: yes—scrolling menus with softkeys
RPN stack roll: bidirectional
Radix mark: selectable ./,
Thousands separator: toggleable
Number formats: FIX/FLOAT
Equation Solver: yes
Absolute value, integer truncation: both
Business/Financial functions:
TVM solvers: standard, odd-period, Canadian, Lunisolar (P/nYR, C/nYR)
TVM shortcuts: n=x*P/YR, i=x/C/YR
Cash flow capacity: 96 groups, unlimited frequency
Cash flow analysis: NPV, NFV, NUS, IRR, MIRR, FMRR, normal/discounted PBP
Bond calculations: all types, price, yield, coupon rate, accrued interest, normal/modified Macaulay duration
Amortization: accumulated interest, balance
Depreciation calculations: Standard/French Straight line, Declining balance with or without crossover, Sum of Years’ digits, French Amortization, US Accelerated Cost Recovery System
Markup calculations: percent of cost and price
Break-even analysis: yes
Black-Scholes equation: yes
Interest rate conversions: yes
Programming features:
Storage: 576 steps
Model:
Editing:
Display: yes
Flow control:
Branch target: addressed by step number or label
Unconditional branching: yes
Conditional branching: any
Subroutines: GSB step number, XEQ label
Loops: any
Addressing: direct and indirect
Flags: yes
As you can see, I am up in the air about whether the calculator should be keystroke programmable or use a more sophisticated, computer-like, programming language and whether it should have one or two shift keys. What do you all think about these issues?
One thing about the key layout of the CW-32B:
Since I conceive it as being fundamentally a programmable calculator, the CW-32B will have a row of keys dedicated to programming functions, that much is undebatable.
Here is what the key row will look like:
Unshifted: P/R|GOTO |LBL |?x=|IS? |SOLVE|FLAGS
Shifted: R/S|GOSUB|RETURN|?≠|DS? |STOP |PAUSE
The menus will be:
GOTO—1 GOTO, 2 GOTOT, 3 GOTOF
Shift+GOSUB—1 GOSUB, 2 GOSUBT, 3 GOSUBF, 4 XEQ
Shift+GOSUB+4—1 XEQ, 2 XEQT, 3 XEQF
?x=—1 ?=, 2 ?≈, 3 ?≥, 4 ?≳, 5 ?≤, 6 ?≲
Shift+?≠—1 ?≠, 2 ?≉, 3 ?<, 4 ?>
IS?—1 ISE, 2 ISA, 3 ISG
Shift+DS?—1 DSE, 2 DSA, 3 DSL
FLAGS—1 SF, 2 CF, 3 TF
FLAGS+1—1 SF, 2 SFT, 3 SFF
FLAGS+2—1 CF, 2 CFT, 3 CFF
FLAGS+3—1 TF, 2 TFT, 3 TFF, 4 TFI
FLAGS+3+4—1 TFIS, 2 TFIST, 3 TFISF, 4 TFIC, 5 TFICT, 6 TFICF
What do you all think of this layout?

[walter b]

Quote:What do you all think of this layout?

I think the layout of your post could really benefit from some formatting, in particular the final part of it.

d:-/

[Marcus von Cube]

(09-29-2014 08:29 PM)Joseph_21sv Wrote:  Memory: 74 numbers [A…Z, a…z, 0…9, .0….9, theta, (I)]+144 numbers addressed by (I)

This seems to be an unnecessary restriction. The 41C and the 34S can live without a dedicated index register. Any register can be used for indirect addressing.

[Paul Dale]

(09-30-2014 08:35 PM)Marcus von Cube Wrote:  This seems to be an unnecessary restriction. The 41C and the 34S can live without a dedicated index register. Any register can be used for indirect addressing.

I'd side with Marcus here. Make every register indirect capable.

I was set on having I and J registers on the 34S to provide easier compatibility with old programs, but there really was no pressing need. If we had had RAM, a full set of lettered registers might have been on the cards too.


Pauli

[Paul Dale]

(09-29-2014 08:29 PM)Joseph_21sv Wrote:  Memory: 74 numbers [A…Z, a…z, 0…9, .0….9, theta, (I)]+144 numbers addressed by (I)

Are you sure you want all these? Some could be tricky to distinguish 0 and O, 1 and l.


Pauli

[Joseph_21sv]

(09-29-2014 08:57 PM)walter b Wrote:  

Quote:What do you all think of this layout?

I think the layout of your post could really benefit from some formatting, in particular the final part of it.

d

Sorry the post doesn't appear to make it particularly clear what I want your opinions on, but it is supposed to be asking about the layout of the programming keys and the program function menus—not its own layout as that is unhelpful to designing the calculator. So what do you all think of the layout of programming keys and the programming menu tree?
Once again, for your reference, here is what the key row and menu tree will look like:
Unshifted: P/R|GOTO |LBL |?x=|IS? |SOLVE|FLAGS
Shifted: R/S|GOSUB|RETURN|?≠|DS? |STOP |PAUSE
The menus will be:
GOTO—1 GOTO, 2 GOTOT, 3 GOTOF
Shift+GOSUB—1 GOSUB, 2 GOSUBT, 3 GOSUBF, 4 XEQ
Shift+GOSUB+4—1 XEQ, 2 XEQT, 3 XEQF
?x=—1 ?=, 2 ?≈, 3 ?≥, 4 ?≳, 5 ?≤, 6 ?≲
Shift+?≠—1 ?≠, 2 ?≉, 3 ?<, 4 ?>
IS?—1 ISE, 2 ISA, 3 ISG
Shift+DS?—1 DSE, 2 DSA, 3 DSL
FLAGS—1 SF, 2 CF, 3 TF
FLAGS+1—1 SF, 2 SFT, 3 SFF
FLAGS+2—1 CF, 2 CFT, 3 CFF
FLAGS+3—1 TF, 2 TFT, 3 TFF, 4 TFI
FLAGS+3+4—1 TFIS, 2 TFIST, 3 TFISF, 4 TFIC, 5 TFICT, 6 TFICF
On a side note, the (I) register is not the only index register of the calculator—it is merely the only dedicated index register—every other register can store indices of other registers, but it is necessary to use a prefix key to flag the value in the register as an index rather than a normal number.

[rprosperi]

(10-01-2014 09:49 PM)Joseph_21sv Wrote:  

(09-29-2014 08:57 PM)walter b Wrote:  I think the layout of your post could really benefit from some formatting, in particular the final part of it.

d

Sorry the post doesn't appear to make it particularly clear what I want your opinions on, but it is supposed to be asking about the layout of the programming keys and the program function menus—not its own layout as that is unhelpful to designing the calculator. So what do you all think of the layout of programming keys and the programming menu tree?
Once again, for your reference, here is what the key row and menu tree will look like:
Unshifted: P/R|GOTO |LBL |?x=|IS? |SOLVE|FLAGS
Shifted: R/S|GOSUB|RETURN|?≠|DS? |STOP |PAUSE
The menus will be:
GOTO—1 GOTO, 2 GOTOT, 3 GOTOF
Shift+GOSUB—1 GOSUB, 2 GOSUBT, 3 GOSUBF, 4 XEQ
Shift+GOSUB+4—1 XEQ, 2 XEQT, 3 XEQF
?x=—1 ?=, 2 ?≈, 3 ?≥, 4 ?≳, 5 ?≤, 6 ?≲
Shift+?≠—1 ?≠, 2 ?≉, 3 ?<, 4 ?>
IS?—1 ISE, 2 ISA, 3 ISG
Shift+DS?—1 DSE, 2 DSA, 3 DSL
FLAGS—1 SF, 2 CF, 3 TF
FLAGS+1—1 SF, 2 SFT, 3 SFF
FLAGS+2—1 CF, 2 CFT, 3 CFF
FLAGS+3—1 TF, 2 TFT, 3 TFF, 4 TFI
FLAGS+3+4—1 TFIS, 2 TFIST, 3 TFISF, 4 TFIC, 5 TFICT, 6 TFICF
On a side note, the (I) register is not the only index register of the calculator—it is merely the only dedicated index register—every other register can store indices of other registers, but it is necessary to use a prefix key to flag the value in the register as an index rather than a normal number.

I believe what was meant (correct me if wrong Walter) is that it is much easier to assess your proposed keyboard layout if it can be viewed as it would be on a real machine, such that there are distinct rows and columns of keys with labels.

One idea is to use a fixed-pitch font (e.g. Courier) and space all rows equally, and maybe a simpler suggestion would be to take a photo of the mock-up you may have made (paper, not physical) and insert said photo.

I've found that the easier you make it for folks to review something, the more likely they are to do so. As laid-out above (and since it was not coming from my mind) it's difficult to mentally visualize the layout you propose.

HTH

[Joseph_21sv]

(10-01-2014 10:00 PM)rprosperi Wrote:  

(10-01-2014 09:49 PM)Joseph_21sv Wrote:  Sorry the post doesn't appear to make it particularly clear what I want your opinions on, but it is supposed to be asking about the layout of the programming keys and the program function menus—not its own layout as that is unhelpful to designing the calculator. So what do you all think of the layout of programming keys and the programming menu tree?
Once again, for your reference, here is what the key row and menu tree will look like:
Unshifted: P/R|GOTO |LBL |?x=|IS? |SOLVE|FLAGS
Shifted: R/S|GOSUB|RETURN|?≠|DS? |STOP |PAUSE
The menus will be:
GOTO—1 GOTO, 2 GOTOT, 3 GOTOF
Shift+GOSUB—1 GOSUB, 2 GOSUBT, 3 GOSUBF, 4 XEQ
Shift+GOSUB+4—1 XEQ, 2 XEQT, 3 XEQF
?x=—1 ?=, 2 ?≈, 3 ?≥, 4 ?≳, 5 ?≤, 6 ?≲
Shift+?≠—1 ?≠, 2 ?≉, 3 ?<, 4 ?>
IS?—1 ISE, 2 ISA, 3 ISG
Shift+DS?—1 DSE, 2 DSA, 3 DSL
FLAGS—1 SF, 2 CF, 3 TF
FLAGS+1—1 SF, 2 SFT, 3 SFF
FLAGS+2—1 CF, 2 CFT, 3 CFF
FLAGS+3—1 TF, 2 TFT, 3 TFF, 4 TFI
FLAGS+3+4—1 TFIS, 2 TFIST, 3 TFISF, 4 TFIC, 5 TFICT, 6 TFICF
On a side note, the (I) register is not the only index register of the calculator—it is merely the only dedicated index register—every other register can store indices of other registers, but it is necessary to use a prefix key to flag the value in the register as an index rather than a normal number.

I believe what was meant (correct me if wrong Walter) is that it is much easier to assess your proposed keyboard layout if it can be viewed as it would be on a real machine, such that there are distinct rows and columns of keys with labels.

One idea is to use a fixed-pitch font (e.g. Courier) and space all rows equally, and maybe a simpler suggestion would be to take a photo of the mock-up you may have made (paper, not physical) and insert said photo.

I've found that the easier you make it for folks to review something, the more likely they are to do so. As laid-out above (and since it was not coming from my mind) it's difficult to mentally visualize the layout you propose.

HTH

Here is the mock-up (pardon the camera if it is not very legible):

[walter b]

(10-02-2014 04:03 AM)Joseph_21sv Wrote:  Here is the mock-up (pardon the camera if it is not very legible):

I'm afraid your image link doesn't really work, at least for me.

d:-?

[Joseph_21sv]

Here is the mock-up once again (pardon the quality as it is not particularly legible):
What do you all think of it?
Once again, on a side note, the (I) register is not the only index register of the calculator—it is merely the only dedicated index register—every other register can store indices of other registers, but it is necessary to use a prefix key to flag the value in the register as an index rather than a normal number. And this is the menu tree:
GOTO—1 GOTO, 2 GOTOT, 3 GOTOF
Shift+GOSUB—1 GOSUB, 2 GOSUBT, 3 GOSUBF, 4+1 XEQ, 4+2 XEQT, 4+3 XEQF
?x=—1 ?=, 2 ?≈, 3 ?≥, 4 ?≳, 5 ?≤, 6 ?≲
Shift+?≠—1 ?≠, 2 ?≉, 3 ?<, 4 ?>
IS?—1 ISE, 2 ISA, 3 ISG
Shift+DS?—1 DSE, 2 DSA, 3 DSL
FLAGS—1+1 SF, 1+2 SFT, 1+3 SFF
2+1 CF, 2+2 CFT, 2+3 CFF
3+1 TF, 3+2 TFT, 3+3 TFF,
3+4+1 TFIS, 3+4+2 TFIST, 3+4+3 TFISF, 3+4+4 TFIC, 3+4+5 TFICT, 3+4+6 TFICF
But, what do you all think of this menu tree: Is it too exhaustive or is it just right?
Moreover, what do you all think of the 32B having a dedicated key for the Black-Scholes equation (one of the selling points of the 30b is that this equation is preprogrammed into its firmware but the keyboard is so small you need to use a shift key to access the Black-Scholes solver which makes that selling point sort of moot—who really cares about it if it is a shifted function?)?