Re: A89: user functions
[Prev][Next][Index][Thread]
Re: A89: user functions
That was just an example, to show the limitations of TI-BASIC =P. You
shouldn't be using an 89 in Cal I anyway. I know that it is a simple process
to do by hand, and I actually wrote a function in BASIC to do it (using
variable substitution, like this:
ImpDiff (v=r+s,{v,r,s},t,1) would convert it to -
v(t)=r(t)+s(t);
then d(r(t),t)=d(r(t),t)+d(s(t),t);
and then so on so forth for different order derivatives.
The problem is that even so simple a function is uggggggglllllly. It's quite
long and fairly slow. First, you have to loop to convert all the dependant
variables to functions of the independant variable (I know you could just
enter them that way, but I didn't want to =P), and then loop for the order
of the derivative. Big, nasty, and slow. To do anything in higher
mathematics (<rant>which the 89 should support anyway, and to those who
whine that it wouldn't be allowed on the SAT should be shot, because you
don't need sybolic manipulation yet</rant>) would be next to impossible (or
horrendously slow) in BASIC. And the screw-up with TI-GCC was my fault
(tried graphing a user function, and the calc crashed =0), but you shouldn't
have to have a kernel installed just to intercept the TIOS's stupid
behavior. I hope that the SDK will have full support for user functions.
> > Say I wanted to write a
> > function for multivariable implicit differentiation...a simple example:
> >
> > v = r+s, find dv/dt,dr/dt,and ds/dt
> >
> > answer: dv/dt = dr/dt+ds/dt (I think that's right,been too long since
> > Cal I =P)
>
>Just a question: Why are you using a function for this? People are
>always underestimating the capabilities of a TI-89. Just write:
>
>d(v(t)=r(t)+s(t),t)
>
>and the result will come out like this:
>
>d(v(t),t)=d(r(t),t)+d(s(t),t)
>
>(for this simple example)
>
>If you want to get more complicated and solve this for something, and
>can't do that yourself - assuming you are one of those usual TI-89 users -
>you
>just have to substitute a variable for what you want to solve for, and
>voilą!
> You can even do that by writing:
>
>ans(1)|d(r(t),t)=a
>
>Now if that didn't help you, try doing this with the SDK, or worse,
>TI-GCC. (I love that EStack - never used it, but it is probably good for
>RPN
>;-).)
>
>Don't worry about the look of these equations; they look a lot nicer with
>Pretty Print. I cannot guarantee that these work with AMS versions other
>than 2.03. For any further questions, mail me or turn this over to the
>math
>list people.
>
>BTW: Scott, NO!
>
>Sebastian Reichelt
>
>--
>Sent through GMX FreeMail - http://www.gmx.net
>
________________________________________________________________________
Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com