Path: utzoo!news-server.csri.toronto.edu!cs.utexas.edu!sun-barr!olivea!uunet!mcsun!i2unix!alessia!ares
From: ares@alessia.dei.unipd.it (Nicola Catacchio 259126)
Newsgroups: comp.sys.handhelds
Subject: Re: HP48: User-defined derivative problems
Summary: other help
Message-ID: <10063@alessia.dei.unipd.it>
Date: 4 Mar 91 16:49:56 GMT
References: <2904833@cc.sfu.ca>
Organization: DEI Padova University
Lines: 25

In article <2904833@cc.sfu.ca>, Dan_Ciarniello@cc.sfu.ca writes:

	During my Signal Theory courses, I had the same problem and I solved
it in this way: since you can't define a function whose value in 0 is +oo
and 0 for any other value, I considered the Delta function in sampled spaces;
it can be considered as a rect whose width is the sampling quantum, 
STOred in a variable (e.g. 't') and whose value the INVerse of t.
	It may be too simple, but when you need to sample a signal expressed
as an algebraic whith some Delta functions in it, simply STOre the width
of sampling interval divided by the number of samples in the 't' variable,
and evaluate the algebraic in every sampling point.
	Define the Heaviside function H(t) as :
'H'
<< -> x 'x>=0' >>
and its derivative as:  
der>=
<< -> x '(x>=0-x>=t)/t' >>

	You can obtain the name 'der>=' first,deriving the algebraic 'H(x)'
and then getting the name OBJ->ing the algebraic resulting.
	With such a definition I got not only the Delta(x) as the derivative
of H(x) but also all the further derivatives of Delta(x).


	Nicola Catacchio (ares@alessia.dei.unipd.it)