<html><head></head><body><strong>Från:
</strong>
rex <rex@nosyntax.net>
<strong></strong>:<br>> I ranted about the waste of time inculcating integration tricks when I was teaching, but my drift into recognizing the relative importance of statistical inference came later. <br><br>I have just finished reading Paul J. Nahin's "Inside Interesting Integrals", which is a romp through evaluating definite integrals. I think as a form of mental puzzle or skilled artform it is great. <br><br>But it is relatively rare that a complex analytic answer illuminates anything: figuring out the proportionality of the answer to the parameters of the problem is often *far* more important than knowing the exact form, and these days numerical evaluation can handle nearly anything (Nahin "tests" his answers by using Matlab quadrature, which usually gives really good results even for pretty tough functions - the cases where it fails are noteworthy). <br><br><br><br><br><br>Anders Sandberg,
Future of Humanity Institute
Philosophy Faculty of Oxford University<br><br></body></html>