Path: utzoo!censor!geac!torsqnt!news-server.csri.toronto.edu!cs.utexas.edu!swrinde!zaphod.mps.ohio-state.edu!wuarchive!emory!gatech!mcnc!uvaarpa!haven!decuac!news
From: baker@wbc.enet.dec.com
Newsgroups: comp.sys.amiga
Subject: RE: A3000UX Competition
Message-ID: <1990Dec12.194125.3150@decuac.dec.com>
Date: 12 Dec 90 19:28:36 GMT
Organization: Digital Equipment Corporation
Lines: 26

-Message-Text-Follows-

	Without getting involved in the argument itself, I'd like
	to mention a relevant article in the December issue of the
	Communications of the ACM.  It's called, "An Empirical Study
	of the Reliability of Unix Utilities", authors are Barton P.
	Miller (et al).

	The gist of the article is that these folks wanted to see
	how robust the utility programs were under random input
	(the idea was inspired by bad phone lines on a rainy night)..

	Testing a number of different Unices, they sent random character
	strings as input to the various utility programs, and on the 
	average, 24% of the programs reacted *very* badly.  "Very badly"
	in this context means that the utility either went into an
	infinite loop, or it core dumped...

	Unfortunately, although SunOS, BSD, and AIX are represented in
	the tests, ATT V.4 isn't.

	So, will *your* next command be one of the 76% that work, or
	the 24% that don't ?

	Regards
	 Art Baker