Path: utzoo!utgpu!watmath!clyde!att!osu-cis!tut.cis.ohio-state.edu!mailrus!cornell!uw-beaver!uw-june!rik
From: rik@june.cs.washington.edu (Rik Littlefield)
Newsgroups: comp.arch
Subject: Re: Professional Programmers and cos(atan(x))
Summary: We've been had...
Message-ID: <6794@june.cs.washington.edu>
Date: 26 Dec 88 20:15:29 GMT
References: <1984@eos.UUCP> <28200245@mcdurb> <321@taux02.UUCP> <32353@think.UUCP> <10790@tekecs.TEK.COM>
Organization: U of Washington, Computer Science, Seattle
Lines: 14

Remember the posting about converting cos(atan(x)) to 1/(1+x*x)?
It touched off quite a chain of postings about what professional 
programmers should and shouldn't be responsible for.  

Well, guys, I'm not sure it was deliberate, but we've been had.  In fact,

  cos(atan(x)) = 1/SQRT(1+x*x)

(Hint: what's the hypotenuse of a right triangle whose sides are 1 and x?)

I hope at least some of you get the same chuckle I did out of this.  I'll
defer comment on the efficiency of the correct substitution to Herman Rubin.

--Rik