# HP Forums

Full Version: Proposed Small Calculator
You're currently viewing a stripped down version of our content. View the full version with proper formatting.

Here are the calculator designs that I want.
Small. Simple. Versatile.

Features

• Algebraic or RPN • Undo
• Fractions (a b/c) • Foot / Inch input (a' b")
• Trigonometry • Logarithms
• Degrees or radians • Deg : Min : Sec input
• 13 memory registers • 6 programs (A-F) [Edited]
• Mortgage calc's • Probability
• Date math • Virtual tape display
• Statistics • Solver
• Complex numerals • Two-dimensional vector
• Decimal, Binary, Octal, or Hexadecimal numerals
• Fixed or significant digits

Edited: 3 Oct 2009, 1:44 p.m. after one or more responses were posted

Small, versatile. Maybe not so simple. (At least to get HP to build one).

I especially like Foot/Inch input (a' b") .

[Edited grammar]

Edited: 2 Oct 2009, 3:09 p.m.

I like the fraction display that doubles as an exponent display.

Don't forget R-P and P-R

No ENG display?

I'd have an unshifted X-Y key.

Dave.

The [a b] key and the a«b key are meant to toggle between rectangular and polar modes.

I've always thought that an ENG mode misses the point. What I want is to be able to specify the number of significant digits. The calculator should display as many digits after the decimal as is necessary to display the number of digits specified by f Sig# __. That number of digits should be determined on the fly and will differ for every engineering or scientific calculation. Examples: 4.03 lb-ft (3 significant digits) or 48.4 lb-inches (again, 3 significant digits).

The foot/inch input may not help in much of the world, but it is a big deal in the USA. I know of at least three companies that sell "construction" calculators. The people I know with one of those calculators all bought them just to get the inch/foot capability.

The "significant digits" seems like a nice option, but ENG is useful, at least in electronics, where you have "gigas", "megas", "kilos", "millis", "micros", "nanos", "picos" and so on, in units of daily usage.

And, by the way, allow for 3 simultaneous equations, and/or 3x3 matrices at least; don't settle for just 2 dimensional vectors.

For BIN, OCT and HEX, a WORD SIZE function is needed. And please, allow for comma and dot radix; thousands separators should be optional (as flags 28 and 29 in the HP 41C).

Best regards!!

those are cool designs. I like Poc 31 better. Better still, remove the Algebraic = from both designs :-)

Quote:
... remove the Algebraic = from both designs :-)

NO! :-(

How is the solver supposed to work?

I don't see any of the usual programming functions and the display doesn't seem suitable for formula entry.

- Pauli

I am an RPN old-timer and I agree with you, Martin. I really like the 27S--an "algebraic" and the 48GX ability to parse "algebraic" expressions.

One of the first programs I ever wrote for my 11c wen I was a kid, was feet-inches-eighths to decimal feet and vice-versa. I think I've posted it here before, and someone even had some elegant refinements!

One of the neat things about making a pr0gram for F-I-E is that you can go finer than F-I-E merely by inputting or returning remainders. I allowed the Decimal Feet to FIE to give the decimal remainder of the eighths--which gives you 16ths etc so you don't even need a different program for F-I-S.

Of course nowadays this is almost exclusively the job of a spreadsheet--the Holy Grail of tabular repetitive calculations!

Edited: 2 Oct 2009, 10:28 p.m.

Quote:
I really like the 27S--an "algebraic" ...

My favorite!
Quote:
... and the 48GX ability to parse "algebraic" expressions.

I have an SX - still finding it difficult to learn, but I keep trying.

Quote:
One of the first programs I ever wrote for my 11c when I was a kid, was feet-inches-eighths to decimal feet and vice-versa. I think I've posted it here before, and someone even had some elegant refinements!

Where is it??

Bill;
That's a real useful program for construction layout here in the states, and from a kid too. Engineers and Surveyors use feet, tenths and hundredths. Architects and Tradespeople use feet, inches and fractions down to eighths. Like most old time layout guys i have every inch (1"=.08, 5"=.42) and the even fractions (6"=.50, 4"=.33) memorized, along with most of them just from repetition; there's only 96 eighths. Since a hundredth is so close to an eighth; one never has to add or subtract more than .04' to translate from decimal to Architectese. It goes quicker than using a calculator to do it in your head or from memory too.
Personally; i was too lazy to write my own program like you did at first. My PPC ROM would do my thinking for me - till rote memorization took over.

Nice looking designs! I'd prefer the 31.

Some questions (asking's more efficient than speculating):

• What is a "virtual tape display"?
• Looks like you're suggesting a 7-segment display. If true then how are exponents handled in the small section reserved for them?
• Looks like fractions are confined to 2 digit denominators -- based on the assumption above.
• What is "odd" for and how shall it work?
• I see a yellow "Int" near to the decimal point -- what does it do? I don't see "Frac". Or shall it mean "Integrate"?
• What do you plan to offer behind "MODE"?
Some suggestions: I'd vote for ENG and a primary x<->y as others did before. + I won't waste a primary space for %.

Ceterum censeo: HP, launch a 43S.

Walter

The small calculator gets a decal on the back. On it goes some seldom-used mode changes. I thought they should be:

g Mode A -> Algebraic -><- RPN

g Mode C -> Comma -><- Dot

g Mode D -> M.D.Y -><- D.M.Y

g Mode E -> Record Macro (in POC 30 only)

g Mode 1 -> inch precision = 1/x

g Mode 2 -> Binary

g Mode 8 -> Octal

g Mode ^ -> Display Statistics

For bit bangers like you, I would be happy to add:

g Mode 3 -> Unsigned -><- 2's complement

g Mode 4 -> Word size = x

Matrices are intentionally excluded. It's a 31 key, one-line, no-menu calculator. I had to leave something out.

Edited: 3 Oct 2009, 10:31 a.m.

To run a program, type gA, or gB, ... , gF.
(When in HEX mode that gets you a hex numeral, so you have to type "g" twice to get the program when in HEX mode.)

Note that RPN users have no use for ( or ), so they get single-keystroke access to programs A and B.

To write a program, you have two options:
1. Simple programs can be recorded from the keyboard. To record program A, one would type g REC A, then keypress your program, then type g REC.
2. More complex programs can be written on a computer and input through a USB cable (sold separately).

The Solve key works the same way it does in the hp15C. Put a program into A, then type f SOLVE A.

A virtual tape display allows me to scroll back through the previous operations with the up arrow and down arrow keys. I personally would use it mostly when I'm paying monthly bills and I want to check my input.

If a seven segment display saves battery life, then I'm for it. Otherwise a good dot matrix display is better.

If an exponent runs more than 2 digits, the calculator can always revert to the old one-line style of exponent display. And you're right about the fractions. Of course, a dot-matrix display would solve those limitations.

The "Int" returns an integer result. With it, you can program your own "Frac" key or your own modulo.

The Modes are covered in a post above.

Regarding ENG and X-Y, perhaps you are right. Opinions vary, and I might come around to your way of thinking. But I'm committed to a small, 31 key calculator. That means something else gets displaced.

Edited: 3 Oct 2009, 10:34 a.m.

The "odd" key is for mortgage calculations. It is a reminder that memory register 0 holds a value for the number of odd days at the beginning of a mortgage term.

Quote:
g Mode ^ -> Display Statistics

??

How do you want to access the necessary operations? And which do you want to offer?

BTW, keys A-F allow for 6 programs only (see your first post).

Hi db,

That is so true--about faster when memorized. I've gotten to know many (though not all) of those 96 eighths by heart: 0.0208333333 = 1/4" , .0625' = 3/4" etc. Some of this from typing in zoom scale 0.02083333xp into autocad to get a viewport to the right scale....

In the old days, when I used my program to go through a table of offsets, I would find that after a while, some of the patterns were memorized very quickly. like 7-7-4 is 7.625, 3-2-4 is 3.208333 etc.

And so in the end, the program was more a memorization aid than a program. Better for the brain that way anyway!

Edited: 3 Oct 2009, 12:38 p.m.

Thanks for catching the error regarding the number of programs. I've edited the original post.

The statistics function is an odd one, I agree. First, a bit of design philosohpy: I think calculators today serve best as a complement to computers, not an alternative to computers. That means, for me, simple one-line calculations get done on a calculator. More complicated work gets done on a computer.

That philosophy drives some decisions on what to include Examples: (1) complex numbers are in, but 3x3 matrices are out. (2) Mortgage calcs are in but NPV is out.

I lost my nerve when it came to accumulated statistics. I really think that simple statistics should be done on a spreadsheet and complicated statistics should be done on SPSS or something similar. You can't check your work if do accumulated statistics on a calculator. So I don't want to give it much keyboard space.

So my, admittedly lame, statistics scheme is:
(1) Use g [sigma] to enter a single datum.
(2) Then type g MODE ^ to enter Statistic Retrieval mode
(3) Then use the up and down arrows to retrieve mean, variance, etc.

Quote:
I've always thought that an ENG mode misses the point. What I want is to be able to specify the number of significant digits. The calculator should display as many digits after the decimal as is necessary to display the number of digits specified by f Sig# __. That number of digits should be determined on the fly and will differ for every engineering or scientific calculation. Examples: 4.03 lb-ft (3 significant digits) or 48.4 lb-inches (again, 3 significant digits).

SCI n and ENG n on HP models are identical, except that the value of the base-10 exponent in ENG is always an integer multiple of three, while SCI produces a mantissa of magnitude greater than unity but less than ten. n is the user-defined number of significant digits to display.

What you may want is for the calculator to determine the valid number of significant digits in a calculation or chain of calculations. I am unaware of any calculator that does that. It seems possible, but a couple things are necessary:

• A means of activating and deactivating "significant digits" mode, so the calculator knows when to apply the process. (Maybe the "sig #" function would do that, but then a display annunciator would be needed.)

• A means of specifying an exact number without having to append zeroes. (It would be a nusiance in "significant digits" mode to see and utilize the full calculated value of 2*pi without having to enter '2.000000000000'.)

The calculator must track the number of significant digits of every result, imposing some extra overhead on the machine. Users, however, generally want it to provide the full-accuracy calculation, without analyzing results to determine what it signifies. In practice, if the operands don't have a decent number of significant figures (at least 4-5), the results often won't be very informative.

Quote:
To run a program, type gA, or gB, ... , gF. (When in HEX mode that gets you a hex numeral, so you have to type "g" twice to get the program when in HEX mode.)

To write a program, you have two options: 1. Simple programs can be recorded from the keyboard. To record program A, one would type g REC A, then keypress your program, then type g REC. 2. More complex programs can be written on a computer and input through a USB cable (sold separately).

Programmability (or entry of equations) is essential for your concept, due to SOLVE. Your proposed 7-segment display doesn't allow for equations.

Why not provide numerical integration along with rootfinding -- a pioneering design objective of the HP-34C of 1979 incorporated into the HP-15C of 1982?

If, e.g., "REC A" essentially does PRGM, then LBL A -- I see no particular means of editing a program once it is entered. Also not provided are the more-advanced HP-15C capabilities such as multiple entry points (via LBL), GOTO, subroutines, looping, conditional tests, and flags.

The capability of editing is essential for useful programming. That of the HP-12C and HP-10C is barely usable, due to the lack of insert/delete editing. I have a 1981 Casio whose built-in Simpson's Rule integration requires programmability. It offers 38 steps amongst two programs; steps are executed as they are entered and recorded (yielding possible math errors); programs are not displayable or editable -- a program must be deleted and re-entered if an error is made.

What would the format of USB-uploaded programs be?

Quote:
Complex numerals

More than just the ability to enter and calculate with complex numbers is needed. They must be displayable, and the germane utility functions -- or a means of doing them -- must be provided.

Quote:
If a seven segment display saves battery life, then I'm for it. Otherwise a good dot matrix display is better.

Dot-matrix displays aren't battery hogs -- time clocks, I/O, and peripherals are. Thin batteries such as the CR2032 also don't seem to last, I assume due to less volume available for chemicals.

Since you seem either to have, or have seen, an HP-15C, why not review this discussion of its functional, organized keyboard layout, and -- better yet -- read the fine Hewlett-Packard Journal article from May 1983, available on CD or DVD ROM from the MoHPC?

-- KS

Edited: 3 Oct 2009, 9:58 p.m. after one or more responses were posted

• Quote:
g Mode D -> M.D.Y -><- D.M.Y

No Y.M.D dates?
Quite a bit of the world uses this as their date format.

- Pauli

Quote:
g Mode E -> Record Macro (in POC 30 only)

Why not in the 31 too?
It has a solve function and would need programmability to support this.

- Pauli

I love the YYYY.M[M].D[D] format. It is unambiguously different than the other two date formats. So it doesn't need to have a mode set in order to be recognized.

In landscape format, I was able to find room for a g REC key. So that calculator doesn't need a REC on the back decal.

You are correct in that no calculator does significant digits the way I want. That doesn't make it right. Also, you don't have to type 2.0000000000. The result will be written to the number of digits previously specified by f Sig#__. So just type 2.

This keyboard does not allow you to edit a program, that's true. You have to start each program from scratch. Moreover, this keyboard does not provide for looping, conditionals, or indirection. It's actually a pretty rotten programming tool. So people will use this keyboard only to record trivially small programs. Complicated programs will need a computer and a USB cord.

I'm not trying to duplicate the functionality of the hp35. I think the high end RPN calculator market is served reasonably well. I'm trying to provide a design for the small RPN calculator. And I'm willing to make some trade offs to get that small size. I've got nothing against root finding and numerical integration, except that they will be used infrequently and probably more easily on www.wolframalpha.com. If a way can be found to include integration without increasing the number of keys, then I'll be for it. Otherwise not.

(NOTE: I was editing my previous post when you posted your response. I made a few more after-the-fact changes, too. You might want to re-read it.)

Quote:
You are correct in that no calculator does significant digits the way I want. That doesn't make it right. Also, you don't have to type 2.0000000000. The result will be written to the number of digits previously specified by f Sig#__. So just type 2.

Then, maybe I'm confused. Specifying a number of significant digits to display is exactly what SCI is for. Many calculators -- not just HP's -- offer it.

Quote:
This keyboard does not allow you to edit a program, that's true. You have to start each program from scratch.

I and others would likely find this unacceptable. We can probably live without the other programming features, as spreadsheet programs with BASIC or other language support can tackle such tasks. (Back in the 1970's and early 1980's, however, the likely alternative to using a programmable calculator was to reserve time on a mainframe to run a Fortran program.)

Quote:
I've got nothing against root finding and numerical integration, except that they will be used infrequently and probably more easily on www.wolframalpha.com.

I wouldn't count on a free website always being available and accessible to handle any computational problem one might have. This particular site is quite recent, anyway.

-- KS

Edited: 3 Oct 2009, 10:00 p.m.

Quote:
I and others would likely find this unacceptable

You can say that again. A programming language without looping, conditionals, and the ability to edit your program is not a programming language at all.

I'd say this calc is DOA if programming is important to you (and it is to me).

Quote:
So my, admittedly lame, statistics scheme is: (1) Use g [sigma] to enter a single datum. (2) Then type g MODE ^ to enter Statistic Retrieval mode (3) Then use the up and down arrows to retrieve mean, variance, etc.

This falls far below the standard of any HP since the HP-45 in 1973. IMHO, the automatic calculation of basic statistics is one of the key features of a pocket calculator. By basic statistics I mean Sigma+, Sigma-, x-bar, s, and the necessary sums for linear regression and so-called forecasting.

At least the 4 functions named first should be readily accessible on the keyboard. Sigma+ should be primary.

Edited: 4 Oct 2009, 5:32 p.m.

Quote:
(Back in the 1970's and early 1980's, however, the likely alternative to using a programmable calculator was to reserve time on a mainframe to run a Fortran program.)

Comments like that make me realize just how well-off I was in those days. Honeywell had a computer network based in Minneapolis which was available anywhere if only one had a telephone line and a Silent 700 or equivalent terminal. It was S-L-O-W, but it had an extended BASIC with double precision mathematics, matrix manipulation commands, a substantial library of programs and, of course, printouts and plots on standard paper. There were terminals scattered throughout our facility. My big disappointment was that I couldn't convince my boss to let me have a terminal at my desk.

Karl and Don don't like this calculator. That's fine. There is room in this world for more than one calculator model. You want the full set of functions and a robust programming interface on the keyboard. There are calculators like that being made. You're lucky.

A small RPN engineering calculator is not being made. Such a calculator would have value, even if it had no programming at all.

I hope now that we can part amicably, you to interest in a certain kind of calculator and me on a different path.

Ron, it's not that I don't like the calculator you have designed. I love the Voyager design, something that fits easily in your shirt pocket. Ever since I began programming in the 70's I have dreamed of an ideal programmer's calculator that would fit in your shirt pocket. The 16c came close, but I want something with text input/output. I am also particularly interested in number theory, so my ideal calculator would have keys for prime factors, GCD, LCM, isprime(), and so on. No one makes a Voyager-sized calculator with those functions, and no one is likely to because the target market would be way too small to justify the expense of manufacturing and marketing such a product.

I thought that the 35s, imperfect though it may be, met the needs of most engineers, although it does not fit in the pocket easily.

I think the ultimate solution to what you want may be the Apple iPhone/iTouch system. I have neither, but I know people are developing software apps for them, including some pretty sophisticated emulations of past HP engineering calculators. Now, everyone has their own idea of what the ideal calculator should be, so what I think would be great would be a "calculator toolkit" to let you drag and drop the keys where you want them, with the legends and functions you want, and size the display with how you want it, and so on, and then be able to put that on the iPhone or iTouch platform. Then you would have what you want. Of course, someone would have to build the construction kit, and I'm sure that would not be trivial.

I don't know what direction HP is going in the future (I forfeited that right when I decided not to go to Ft. Collins this year). Maybe they are planning something close to what you describe, but I doubt it.

In any event, I do wish you luck in getting the calc you want. For me, I'm still happy with my 12c in my pocket and 99 lines of code (and unfriendly programming environment too!).

Don

The most cogent reply in this whole thread is the first one. You and Martin Pinckney both think it unlikely that anyone will build this calculator. You're both right.

I've come to agree with Karl about ENG and SCI displays. They should be in the mix. But a decent programming environment needs a lot of keys.

So does a interest in number theory mean that you need complex number input? (For Riemann type stuff?)

Quote:
So does a interest in number theory mean that you need complex number input? (For Riemann type stuff?)
For me personally, no. I guess I'm not into number theory enough to have to deal with that.

Quote:
But a decent programming environment needs a lot of keys.
A macro recorder would be perfectly ok for me. I wish the 20b had one ;^).

Nice design, Ron.