Author 
Comment/Response 
Otje Fauzé

09/13/03 2:50pm
I have to compute the convolution of the functions R > R
f: 0 if Abs[x]=>1
1 if Abs[x]<=1
g: 0 if x <= 0 or if x => 1
1 if 0 < x < 1
Trying something like:
Plot[NIntegrate[g[x  z], {z, 1, 1}], {x, 3, 3}]
works only after a lot of errors, while a simple Riemannalike integration like
\[Epsilon] = 1/1000
Plot[Sum[g[x  z] \[Epsilon], {z, 1, 1, \[Epsilon]}], {x, 3, 3}];
does work. I am using Mathematica 4.2. My questions:
1) Is there a way to compute such convolution symbolical (or by hand?). I can only guess how the convolution would look like, but I'd rather work out the integrals.
2) Why does NIntegrate give so much errors and takes such a long time?
Much thanks in advance,
Otje Fauzé
URL: , 
