In Response To 'Re: Problem with NIntegrate'
I am attaching a file here that shows the problem in the last 3 lines. The function integrated is very complicated. Within a region the function may even have singularities. Nevertheless, outside a boundary region it is defined to be 0. When the limits of integration are change outside this boundary, the results from NIntegrate still change.
I will send a copy of this file by email to wolfram support. Posting here if anyone can figure out the problem.
Attachment: 44ums.nb, URL: ,