Yup, I didn't intend to. RAM areas could be emulated, but patching a copy of TI-OS' page 01 would be too much work. I'm also sticking all the RAM areas that get used into the upper end of RAM (just below the hardware and allocation stacks), allowing programs to avoid the C000h execution limit a little longer.

Great then!

A parser included. It would be nice to not have to follow every forum to see what is included in what. Maybe like putting that in the Sourceforge page?

Some ideas about the parser:

Ability to input manually the hex code for the token

30h: 0 - so the numbers fit their ASCII equivalent
31h-39h: 1-9 - see higher
41h-5Ah: A-Z - The upper font letters, which can be used as other things
and they fit with their ASCII equivalent
10h: Load(arg1,arg2) - loads equation arg1 into variable arg2

FFh: first byte of 2 byte token - don't find where I got that idea, definitely not TI-OS

But as we are close to ASM, why not use real asm things:

12h: ListE(name,element,element_size) - returns element from list name of element size X, treats variable name as a list
11h: CreateVariable(name,size) - creates out-of-the-program variable name
13h: MatrixE(name,row,column,element_size) - returns element from matrix name of element size X, treats variable name as a matrix

[Edit]Does one know about the tools necessary to do OS development?

Benjamin Moody's ostools package and Latenite should be sufficient.
I recommend you use latenite, just because that's what I've been using, and syntax differences can be murder to resolve.

Needless to say, I'll have to withhold judgment until I see how this turns out, but good luck! Hope you see it through to completion.

I'd like to see what becomes of this.

MLC 68K is going well, but I don't really know how MLC for the TI-83+ is going; Madskillz from RevSoft is working on it. Maybe you could share some of those malloc routines with him...eh?