Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!allegra!mit-eddie!think!harvard!seismo!mcvax!boring!steven
From: steven@boring.uucp (Steven Pemberton)
Newsgroups: net.lang
Subject: Re: bugs in assembler vs. C
Message-ID: <6824@boring.UUCP>
Date: Wed, 12-Mar-86 07:41:32 EST
Article-I.D.: boring.6824
Posted: Wed Mar 12 07:41:32 1986
Date-Received: Sat, 15-Mar-86 18:47:48 EST
References: <1437@wucec2.UUCP>
Reply-To: steven@mcvax.UUCP (Steven Pemberton)
Distribution: net
Organization: CWI, Amsterdam
Lines: 16
Apparently-To: rnews@mcvax

In article <1437@wucec2.UUCP> jdz@wucec2.UUCP (Jason D. Zions) writes:
> >> On point B:, studies have shown that the number of errors or changes is
> >> exponentially related to the number of lines of code.  One study indicates
> >> that the average is one error per 1000 lines.
> >You don't mean "exponentially," do you?  (Though it applies to _some_
> >programmers I know....)
> 
> Uh, yes, He did mean exponentially. Really. I don't recall the exact study which
> made this claim, but I do recall that it was exponential in some constant
> fraction of the number of lines.
You can find some details in 'The Mythical Man Month' by FP Brookes, Addison
Wesley, 1975. He quotes several large scale projects, and says that the
exponent is 1.5 of the number of lines of code, regardless of language. It's
a great book, and worth reading for his advice on running large-scale
software projects.
Steven Pemberton, CWi, Amsterdam.