[Serious - programming] I need ideas for program features

I'm writing a command line app that solves mathematical equations.

I need you guys to tell me what kind of math operators you would like to have in such a program. I've already covered most of the obvious ones (plus, minus, divide etc.)...

Code:

---


Supported Operators (token names only, use -L switch to get verbose info):
------------------------------------------------------------------------------------
+      -      *      /      <      >      ==      !=      <=      >=
<<      >>      neg    int    abs    sign    mod    pow    sin    cos
tan    ln      log    exp    asin    acos    atan    sqrt    &&      ||
^^      ~~      imp    eqv    min    max    rint    rand    sum    prod
gcd    mean    stdv    dump    count  swap    swlt    swgt    fact    ?
npr    ncr

---

If any of you boffins can see anything that I've missed out on this list (that you would expect / want / think is cool) then please let me know, I want this to be a good all-round math tool.

Any suggestions will get you a mention in the credits. I'm probably going to put this project up on sourceforge at some point (its ansi C++ so it'll compile on any platform :D)

Please keep it serious. Thanks.

What about log and ln?

good points. I'll add "exp" too while i'm at it

edit: Ln(), Log() and Exp() done :)

How about Mod?

bitwise operators?

Mod() is in progress. Since I need to add an integer version and a floating point version.

There are some bitwise operators already (see the list) but more are on my todo list.

How about a command line that executes a shell command to open IE with Google calculator as the home page?

http://www.xnumber.com/google_calc.htm

nPr and nCr?

Binomial coefficients?

Quote:

---

Originally Posted by zaza

nPr and nCr?

---

Excellent ideas.

IsPrime()?

See first post again to see the latest list. (I'll keep updating it as I add operators).

IsPrime() is a nice idea in theory, but for large numbers that would rapidly become slow. I'll probably put it in though.

Awesome I've got 50 operators now! I want to make 100 if I can.

Keep the ideas coming.

Code:

---

Financial Functions
        DDB        ::        [Double Declining Balance]
        FV        ::        [Future Value]
        Ipmt        ::        [Interest Payment]
        IRR        ::        [Internal Rate of Return]
        Mirr        ::        [Modified Internal Rate of Return]
        NPer        ::        [Number of Periods]
        PMT        ::        [Payment]
        PPMT        ::        [Principal Payment]
        PV        ::        [Present Value]
        Rate        ::        [Interest Rate Per Period]
        SLN        ::        [Straight Line Depreciation]
        SYD        ::        [Sum-of-Years Digits Depreciation]

Other       
        Hex        ::        [Hexadecimal]
        Oct        ::        [Octal]
        Rnd        ::        [Randomize]
        Round        ::        [Rounds a floating point number]

---

IsOdd()/IsEven()

Quote:

---

Originally Posted by sevenhalo

IsOdd()/IsEven()

---

Maybe later. as this can be done with bitwise operators (which I admittedly haven't finished yet).

IsOdd = (mynumber & 1);

How about...

BCD conversion
Binary rotate


Any good?

I've thought about ROL and ROR (rotate left and rotate right) as per IA-32 assembler. however it will require a second operand to declare the width of the integral value being rotated.

No plans for BCD at the moment though. Its an arbitrary numeric encoding and as such doesn't fit the standard DAL/RPN model. maybe a future release

imaginary number support:
square root(negative i) = 0.707106781 - 0.707106781 i

Reals only for the time being. Complex numbers are definitely on the agenda for 2.0 though.

SumOfAll();

Like if you put in 13542
You get: 15

My ideas are only going to get worse...

Did that one a couple of days ago. Same with mean(), prod(), stdv() and so on.

∫ and ∂

...Have fun with those!

Seriously though, sinh, cosh, tanh, sec, csc, cot, sech, csch, coth

Uniary RangeSum(Low,High)

RangeSum(Low,High,Sequence())

RangeSum(Low,High,NumberSet())

CountOfNumbersInRange(Low,High,NumberSet())

Given 3 points...

And a 4th...

Is the 4th inside, or outside the 3?

All functions in Calculous.

I hate it.

Given a set of N 2 dimensional points...SetOfPoints(N,1)

if p3 = F(p1,p2, cw, SetOfPoints)

Whats the next point(P3) clockwise from the 2nd point, nearest to the 2nd point, as you revolve from the first point?

CounterClockwise??? (p3 = F(p1,p2,ccw,SetOfPoints))

Quote:

---

Originally Posted by wossname

I've thought about ROL and ROR (rotate left and rotate right) as per IA-32 assembler.

---

I think you should add LOL and ROFL. Those would be pretty funny functions.

On a serious note, what do you have for constants? Pi and e are obvious ones, but are you thinking of other constants? I'm not quite sure how this will be used in the end, but for some types of practical applications, there are a variety of constants that could be of some value.

As an example, you talk about a math "solver". Does that mean that people might use it as some kind of glorified calculator for typed data? Would the conversion of inches to centimeters ever be of value?

Quote:

---

Originally Posted by Shaggy Hiker

... Would the conversion of inches to centimeters ever be of value?

---

mArr() = CMYK(R,G,B)

mArr2() = RGB(C,M,Y,K)

Of course, for one of those, {Don't ask me ATM which,} you'll need another factor to input, to guide the desired "doping"

[is it 1:1 from cmyk to rgb? I think so..., but it's 1 to many from rgb to cmyk...

Definite integration, indefinite would be rather hard. Then derivation would be nice =)

Quote:

---

Originally Posted by dsheller

Definite integration, indefinite would be rather hard. Then derivation would be nice =)

---

I stated those already!

Quote:

---

Originally Posted by Shaggy Hiker

I think you should add LOL and ROFL. Those would be pretty funny functions.

On a serious note, what do you have for constants? Pi and e are obvious ones, but are you thinking of other constants? I'm not quite sure how this will be used in the end, but for some types of practical applications, there are a variety of constants that could be of some value.

---

Well I'm loading constants and variables from a text file so the user can use his own constants that suit his needs. This also decreases the number of searches that need to be done and increases efficiency.

Quote:

---

Originally Posted by Shaggy Hiker

As an example, you talk about a math "solver". Does that mean that people might use it as some kind of glorified calculator for typed data? Would the conversion of inches to centimeters ever be of value?

---

"typed data" is certainly the simplest way to use a command line app but you can also redirect the output of other programs into my program. A simple example:

Code:

---

./rpnsolve "`cat rpnsolver.cpp | wc` mean"

---

cat reads rpnsolver.cpp and pushes all the data through wc (word count) which returns a list of 3 numbers. This list is then fed into rpnsolve (my program) and its mean() function is called to take the average of those 3 numbers. The final result is (1133 + 3312 + 22369) / 3 == 8938.

Any program that outputs numeric data (either decimal or hex) to the console or to a file can be piped through my program and the data can be operated on.

Thats the general idea :)

imp and eqv.

change between different bases... (binary, oct, and hex at first)

Something to convert decimal to fraction?

Using x^2? (the ^ bit)

Degrees, Radian?

Quote:

---

Originally Posted by kregg

Using x^2? (the ^ bit)

---

thats what the pow() function's for. ("^" has its own meaning in boolean/bitwise operations)

Quote:

---

Originally Posted by kregg

Degrees, Radian?

---

Done:
dxr() degrees to radians
rxd() radians to degrees

everyone: See the first post again.

Split into prime factors?

Quadratic equation solver (given coefficients)?
Cubic....? :)

Ok, I need some more inspiration guys...

I've finally added support for negative numeric literals (the last major omission that I had thus far).

I'm planning to add "stored subroutine" support (loads user-defined math equations from a file for later use) which promises to be awesome :D

But I still need more ideas, the cooler the better. Impress me.......

finding roots for "y = 0.6x² + -4x + -3"
using:
http://mathworld.wolfram.com/images/.../equation5.gif (pic courtesy of Wolfram)

0.6, -4 and -3 are stored into variables "a", "b" and "c" respectively through the config file...

Code:

---

./rpnsolve "=a b c dump"
dump:
[0] == 0.6
[1] == -4
[2] == -3
...
./rpnsolve "=b neg b b * 4 a c * * - sqrt - 2 a * /"
=-0.680532

./rpnsolve "=b neg b b * 4 a c * * - sqrt + 2 a * /"
=7.3472

---

The left-hand root is given in blue and the right hand one in green :D

Hurrah.

For once, I have not been mocked at for giving a suggestion.

Hurrah.

EDIT: I did not see woss's post above before I posted this - Coincidence for Hurrah :sick:

Moved. Please don't post technical questions in the General Discussion / Chit Chat forum.