Path: utzoo!attcan!uunet!tut.cis.ohio-state.edu!ucbvax!agate!stew.ssl.berkeley.edu!johnf
From: johnf@stew.ssl.berkeley.edu (John Flanagan)
Newsgroups: comp.sys.amiga
Subject: Re: C Routines
Message-ID: <1990Jul3.221848.21370@agate.berkeley.edu>
Date: 3 Jul 90 22:18:48 GMT
References: <138087@sun.Eng.Sun.COM> <1858@trlluna.trl.oz> <1990Jul3.193850.15879@midway.uchicago.edu>
Sender: usenet@agate.berkeley.edu (USENET Administrator;;;;ZU44)
Organization: University of California, Berkeley
Lines: 21

In article <1990Jul3.193850.15879@midway.uchicago.edu> arxt@midway.uchicago.edu (patrick  palmer) writes:
>I think you are refering to "Numerrical Recipies in C" by Press, Flannery, 
>Teukolsky, and Vetterling.  It is a Cambridge U. Press book, and has a 
>companion diskette with over 200 subroutines.

Better yet, get the book and ignore the diskette.  There are a few problems
with the Press library C code.  For starters, it is all in single-precision,
and most of the fitting routines do not work very well without being 
converted to double.  Also, he does some stupid (and illegal) things with
array pointers.  In general, the problem with his C code is that it is
basically converted Fortran code;  his Fortran version is quite solid.
The book is wonderful, though, for its thorough explanation of the algorithms.
For best results, get the book, and write your own code using his as a guide.

John


John Flanagan				Center for EUV Astrophysics
johnf@ssl.berkeley.edu			University of California
(...!ucbvax!soc1.ssl!johnf)		Berkeley, CA 94720
Manners Maketh Man.			(415) 643-6308