Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!caen!spool.mu.edu!uunet!lll-winken!ames!xn.ll.mit.edu!ll.mit.edu!news
From: wjc@llex.ll.mit.edu ( Bill Chiarchiaro)
Newsgroups: comp.sys.handhelds
Subject: Numerical Integration on the HP48
Message-ID: <1991Jun5.192839.4847@ll.mit.edu>
Date: 5 Jun 91 19:28:39 GMT
Article-I.D.: ll.1991Jun5.192839.4847
Sender: news@ll.mit.edu
Organization: MIT Lincoln Laboratory
Lines: 32


I have a version E HP48SX and have been trying to get it to
numerically integrate a program object (rather than an algebraic).  I
used to be able to do this on my HP28S, and it was actually useful.
Integration of a program was faster than integration of the
corresponding algebraic expression.  Also, if the definition of the
integrand changed over the range of the variable of integration, then
that could be handled simply by a conditional.  The algebraic object
equivalent would be to split the integration into two or more
integrations and to combine the results.

If I put limits of integration in stack levels 3 and 4, a program
object in level 2, and a quoted name in level 1, I get a "Bad Argument
Type" error when I hit the integral key.  This is in spite of the fact
that the USAG program claims that the integrand can be of any type.  I
have tried storing the program in a variable and putting the
variable's quoted name in level 2.  I tried this in combination with
various settings of flags -2 and -3 and with the specified variable of
integration either existing or not existing as a global variable.  In
some cases, the 48 did not complain and was clearly seeing the program
object, but it never gave a correct result.

I haven't found an answer in the manuals, and the HP Calculator
Support line has not been helpful.

Any ideas?


Thanks,

Bill Chiarchiaro
wjc@ll.mit.edu