Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!usc!ucsd!rutgers!orstcs!jacobs.CS.ORST.EDU!louxj
From: louxj@jacobs.CS.ORST.EDU (John W. Loux)
Newsgroups: comp.sys.handhelds
Subject: Re: Question - Indefinite Integration
Message-ID: <20596@orstcs.CS.ORST.EDU>
Date: 28 Sep 90 17:03:03 GMT
Sender: usenet@orstcs.CS.ORST.EDU
Organization: Oregon State University - CS - Corvallis
Lines: 46

>Ok, forgive the novice question, but we all started with our
>first HP at one point...  Is there a way to evaluate
>an integral sans limits on the 48SX?

I have no problem with a novice question, it's the novice answers that bother
me, especially since the answer has been posted once before.

First off, simply supplying 0 for one of the limits in definite integration
does not work because all functions of zero do not return zero --- and some
will burp and die on you.

Second off, yes it is possible to do indefinite integration using the 48, but
it's not particularly direct or documented.

Consider:  S X dX --- i.e., the integral of X with respect to X.

First, create the expression using arbitrary, symbollic limits:

		S(A,B,X,X)

EVAL'uate this to get the first integration step:

		'X^2/2|(X=B)-(X^2/2|(X-A))'

Split this expression using OBJ-> and drop the bottom three levels:

		'X^2/2|(X=B)'

You can now either evaluate this expression to replace X with B or use OBJ->
again to split the expression and drop the bottom three stack levels again.

The result is either 'B^2/2' or 'X^2/2'.  You can substitute X for B in the
first expression in one of several ways (viz., 'X' 'B' STO EVAL) and add 'C'
if you like ('C' +).

This is the general procedure for expressions that the 48 can integrate
symbollically.

While we're at it, I'm still waiting for someone to respond to the posting on
how one goes about symbollically integrating e^((y^2)/2) symbollically.  The
numeric solution is straight-forward: simply supply numeric limits, build the
expression and use ->NUM to get the numeric result.

I hope I've helped some,

John