Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!cs.utexas.edu!sm.unisys.com!ucla-cs!math.ucla.edu!tony
From: tony@joshua.math.ucla.edu
Newsgroups: comp.lang.c
Subject: The pow() function...
Message-ID: <404@sunset.MATH.UCLA.EDU>
Date: 1 Feb 89 21:58:03 GMT
Sender: news@MATH.UCLA.EDU
Reply-To: tony@MATH.UCLA.EDU ()
Distribution: na
Organization: UCLA Mathematics Department
Lines: 36

Thank you netland for all of the replies regarding my question about
the pow() function.

I received a lot fo replies regarding how inefficient the pow() function
was for flipping signs back and forth.  I realize that this is
true but the only reason I wanted to do this was because of the following:

I am taking the n derivatives of a vector of cos functions

S [w]      =     [1,  cos (w),       cos (2w), ... ,      cos (nw)]
S [w] d/dw =     [0, -sin (w),  -2   sin (2w), ... , -n   sin (nw)]
   .              .    .              .                 .
   .              .    .              .                 .
   .              .    .              .                 .
S [w] d^m/dw^m = [0, -sin (w),  -2^m sin (2w), ... , -n^m sin (nw)]

The closed for solution being :

(-1)^m * (n^m) * cos (nm - mPI/2)

Since I am still testing my program I wanted to use something that
I (and my non-programming engineering friend) can see right off as 
being the closed form solution.

I therefore apologize for mis-using the word elegant (I seemed to have offended
a few people) and I also apologize for posting a simplistic example  (Please no 
more personal attacks!).

I promise all of you people who replied that I will stop using the pow ()
as soon as everything is debugged (and to optimize the heck out of the 
program)!

Thanks again for replying in the first place.


Tony