Just wondering, why run Linux 2.6 if you can run 3.X?

Also, if possible, native support for (Micro)SD cards would be awesome.

Ah okay. Anyway, very nice project. I hope it succeeds.
I'm afraid that I won't be able to use it though. Pretty sure our school only allows TI-83+ and TI-84+ BE/SE calcs

-- EDIT: Also, I'd suggest putting a replaceable Li-Ion or Li-Pol battery in there. Pretty sure most people wouldn't mind plugging their calculators into their chargers like once a week.
-- EDIT2: Is this the CPU? http://www.armadeus.com/english/products-processor_boards-apf28.html

But before evrything, it would be awesome to design a web site for sharing our work, roadmaps, ..., and find other people.

& Coordinate work.

Have you gotten anyone to build the website?  I think that would be a great idea.

Would there be separate sections for French and English?

Have you thought about porting a CAS instead of making one from scratch or would this be harder to do?

The web site is focused on the project only. So nothing on ti, casio, or anything else not related to the project. It's gonna be primary english because it's the internationnal language for working. But evry translators will be welcome, to translate each messages of the site, including forums.

why dev a new CAS engine and not porting one ?

I think to build an engine that is more powerfull and more engeneer oriented. There is some lack of math definition in Ti-BASIC, and it's different from xcas engine.

Xcas is not embded. And if you see the scratch of Xcas, it's 150k lines of code.

So it's not the see the drink.

And you'll be able to use the math libs in C++. these libs will be as advanced ti-basic.
Just think what is possible to do with that ... C++ power with Ti-basic style.

And i want do some functionnality that doesn't exist either in xcas and ti-basic.

And this main advantage, is, you need a specific function for your specific work,
you can code it using the primary blocks of the math engine easyly and ask to
integrate it in futur dev. And if there is no primary blocks for your function,
just integrate yourself the dev team and start the code for it.

i've my own server, and i'll have optical fiber in a few weeks, and i've installed successfully SVN.

And nobody until now started the code of the site. So i'm open to any coders.

I think don't understand my meaning.

C++ is hard, because it lack of advanced class, like expr in Ti-basic.

The math lib goal is to define this with C++ class.

so your code will be :

#include <advancedmathlibs.h>

int main (void)
{

       expr   var1; // create a var of type expr named var1.
       var1 = "integrate(x^2+x+c,x,1,50)"; // set an expression in var1
       std::cout << var1.evaluate() << sdt::endl; // send to std output the evaluation of the the expr

}

you'll able to do var1 * var2, var1+var2, ... and anything that is possible. ( even if it don't do a result in that case you get back the error ).

so you'll have speed of C++ & the power of advanced mathematics coding.

This sounds like a cool project!

I could probably help some out with C++ and webdesign (i know php mysql javascript (html css))
I also know some lua.

And i think too that some BASIC dialect for it would be cool.

It might be nice to have a basic like language though so non-programmers can put basic things into their calculators. I'm not sure I fully understand your reasoning for not using a prebuilt CAS. I'd think something like maxima or octave would work well, and if you want to be able to have it more integrated into C++ it shouldn't be too difficult to build classes on the function they already have. I think for this you should use C/++ as the asm like language of the 84 and have something(python, lua, basic) as a language for quick programs.

It's possible to do a "basic" like language for starters. It's simple to do, just exec in math engine all lines in order. A litle C++ code for running "basic" files will be realy easy to do. The main problem for maxima and others, they don't have 2 major functions that i look for. You can't add it, it's primary blocks.

@Lancelot_du_lac do you want to make C++ a sort of script language ? If so, I recommend to use libgcc for on-the-fly compilation, so you'll have maximal speed.

Also, I think that it should be cool to use Python as basic language, and allow the use of import, so the user'll can use external libs such as PyGame or PyGTK, if we success in porting them (if there really is a need to port something).

C++ will be compiled. so it's not a scripting language, just a native C++ with super charged class.

This sounds like a cool project!

I could probably help some out with C++ and webdesign (i know php mysql javascript (html css))
I also know some lua.

And i think too that some BASIC dialect for it would be cool.

Ready to start a webdesign ?

would be awesome. 

i have a PC with ubuntu SVN and apache2 ready ( already 24h/24 running ), and in 2-3 weeks i'll have optical fiber. ( Already have a domain name, and some crapy stuff on my web server until now but i'll buy a second domain name for this website ).

Just one thing: Will it have a headphone jack? I'd like to write a chiptunetracker/mod tracker for it
Also: will it feature a nice left-handed d-pad? (oh wait it's supposed to be a calculator?)

HD Audio will be available. ( and could be used like a low frequency synthetizer for electric experiments )
d-pad, not really. ( but there is usb host. )
There is azerty/querty keyboard by default, at this step of the project. Packaging, is not easy task, so we'll see.

And yes the primary is to aim be a good calculator.
But we'll never set limitation to this device, for any other kind of workload.