24 replies to this topic

[kucalc]

EDIT: If you want to discuss about a possible CAS for the fx-9860G, then look here: SYMBOLIX CAS

[vanhoa]

* lcm and gcd - supported
* expand - binomials in the works right now
* simplify - radicals in the works right now
* solve - not supported as of right now
* factor - not supported as of right now

=))

[kucalc]

Heh, yeah, I should get more done.

[vanhoa]

The only thing I want to tell u is "CAS is really difficult".
Before I tried to build a simple CAS, and it can do some stuffs as well, althought I have enouph math knowleage to build a simple CAS but TO BUILD A COMPLETE CAS you must VERY GOOD AT MATH (the programming skill isnt the all)!

[kucalc]

I know building a complete CAS will be very difficult, but I want to see how far I can go into developing a CAS system. Also, the fx-9860 community really needs CAS right now. Maybe then the fx-9860 can join the ranks of the big boys like the ClassPad 300 and TI-89...

[Manuel Naranjo]

Cool I will try it, don't worry I don't feel angry or something .
I just thought that having a base is better.
Cheers

[eew]

The only thing I want to tell u is "CAS is really difficult".
Before I tried to build a simple CAS, and it can do some stuffs as well, althought I have enouph math knowleage to build a simple CAS but TO BUILD A COMPLETE CAS you must VERY GOOD AT MATH (the programming skill isnt the all)!

I don't think there are many people who really need it, because I think most people who have a 9860 are not allowed to use such tools, it may be a gread toy.

[Menno]

Hey kucalc, great job

CAS will be a great add-in when it is finished, i guess the demand could be bigger then you'd expect. The 9860 is not used in schools only. I for my self use them professionaly and my boss does not care how i do my calculations, only results count

Menno

[Manuel Naranjo]

I don't think there are many people who really need it, because I think most people who have a 9860 are not allowed to use such tools, it may be a gread toy.

What makes you think that? I'm a EE student and I can use a CAS system. I can't use a Handheld that's why I didn't bought the ClassPad 300 (it looks like a Palm). And I didn't bought a FX2.0 because I thought it was too old and expensive.

Of course you can't use a CAS in the first classes of mathematical analysis, but you can use it when you are doing stuff like circuit theory.

Cheers,
Manuel

[Horrowind]

I think a CAS would be useful
(the calc is allowed but it has CAS... hehehe)

[kucalc]

CAS will be a great add-in when it is finished, i guess the demand could be bigger then you'd expect. The 9860 is not used in schools only. I for my self use them professionaly and my boss does not care how i do my calculations, only results count

That's cool.

I think a CAS would be useful
(the calc is allowed but it has CAS... hehehe)

A CAS would be absolutely useful and the funny thing is that teachers think that there can be no CAS on the fx-9860. :lol

[Horrowind]

how far are you?

[kucalc]

how far are you?

My CAS can now solve (SOLVE command) roots (even complex roots) of polynomials if given the coefficients and the degree of the equation. Now what's really bugging me is how to parse a equation. I need a way to split a equation into terms and also rearrange the equation to make it equal to zero on one side.

For example:

x^3 - 19x - 30 = 0

Plugging in the coefficients and giving the degree of the equation into my CAS, the roots of the equation are:

x = -3
x = -2
x = 5

Yeah, so the solver works correctly, but I want CAS to parse the equation instead of having the user have to rearrange the equation themselves, and then typing in the coefficients and the degree. If a user types in:

x^3 - 19x = 30

The CAS sees it as a string of characters (ex.):

char equa[] = {'x','^','3','-','1','9','x','=','3','0',0};

Does anyone have a idea of how to extract the terms and coefficients from this string? A friend of mine suggested using RPN and a stack system to parse the equation. I would have to convert this equation from infix notation to postfix then. My friend said he will try to write a equation parser for me over the weekend, but I don't think RPN will really help me. Does anyone have any good ideas?

If I can split a equation into it's terms than I can finish implementing EXPAND, SIMPLIFY and SOLVE and that would be a great leap!

[vanhoa]

I think my source should be useful... If you want i will upload it.

About your degree 3 equation, i think you solve it by a pre-input fomula (Cac-na-do - I call his name like that in my contry). Equation degree 4 is still easy. How about 5 or higher? I dont think you know the methol to solve them in algebra (it is not a easy methol) but you may solve it numericaly (a little easy).

[kucalc]

I think my source should be useful... If you want i will upload it.

You could email me: kucalc AT gmail.com or kucalc AT hotmail.com.

About your degree 3 equation, i think you solve it by a pre-input fomula (Cac-na-do - I call his name like that in my contry). Equation degree 4 is still easy. How about 5 or higher? I dont think you know the methol to solve them in algebra (it is not a easy methol) but you may solve it numericaly (a little easy).

I'm not using some standard simple technique for solving polynomials. My polynomial solver is based on some Fortran code of an ACM algorithm for solving polynomials. That 3 degree polynomial was just an example. 5 degree? Mine can solve 10 degree numerically as long as the coefficients are real and give the complex roots as well.

[vanhoa]

I see... That's numerically right?

[kucalc]

Mine can solve 10 degree numerically as long as the coefficients are real and give the complex roots as well.

I see... That's numerically right?

Yes, I said numerically. I have no idea on how to do it algebraically, but my polynomial solver is still so much better than the fx-9860, as the fx-9860 only solves up to 3rd degree.

[Manuel Naranjo]

As far as I know you have two ways to parse this kind of inputs.
One is the one your friend suggested (RPN) the other one (YACC/Lex uses it) is a parse tree. I prefer Parsing Trees, but that request a lit more of memory.

I'm following this course: http://members.tripo...img/Parser.html I'll try to write a simple mathematical parser for the 9860 SDK. I'll do it with Yacc so it's easy to extend.

Also check this site for pieces of code http://www.codecogs.com/

[uvoK]

The link doesn't work.... can I still download the program?

[PierrotLL]

Alexis and I have writen an addin which can differentiate, visit http://www.jeuxcasio...S-graph-85.html to download it.

[uvoK]

Alexis and I have writen an addin which can differentiate, visit http://www.jeuxcasio...S-graph-85.html to download it.

is it for casio fx-9860?the page says it is got "graph 85"?