|  | >	... Some of this stuff just is not that difficult.  
Well, sure, SOME of it is not that difficult. But when you start combining 
things the difficulty ramps up quite fast.
>	... 						    You can
>    actually do some pretty useful stuff in a couple hundred lines of
>    prolog source code.  I'm tempted to buy TURBO prolog and type in the
>    integral tables from the CRC.  
Historical note: The SNOBOL language was designed to solve problems in this 
space, and there have been some pretty good differentiators etc. built with 
it. That turned out not to be the problem, which in the time it was first 
used was mostly the lack of good algorithms, but partly the exponential 
growth in complexity with the size of the problem.
For an example, try solving a 4th-degree polynomial with MAPLE or MACSYMA
etc. There is a closed-form solution, but it takes many pages to print it 
because of the number and complexity of terms, a couple of orders of 
magnitude longer than the solutions for quadratics.
I wish you good luck and great insight. Maybe you'll come up with something 
cheap and useful. (Have you looked at the HP whizbang pocket calculator?)
If you do, let us know.
 | 
|  |     Don't know exactly what you're looking for, but for some VERY
    affordable math software, you might want to check out some of the
    MS/DOS SHAREWARE software.  There is a sampling of what is available in
    TIXEL""::DUA1:[MSDOS.STATISTICS].  Or if you know exactly what you're
    looking for, at least in terms of capabilities, you might want to post
    a note in the TIXEL::IBMPC_SHAREWARE conference.  Press KP7 to add that
    conference to your notebook.
    
    Good luck
    Jon
 | 
|  | The MAPLE people have announced, through Brooks-Cole publishers, an MS-DOS 
"student" version of MapleV for $99+taxes & handling. It has all the
features of the VMS version except some of the packages, and the graphics
and other features are great. See the MAPLE conference for details. 
 |