Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10 5/3/83; site gatech.UUCP
Path: utzoo!linus!decvax!genrad!grkermit!masscomp!clyde!akgua!gatech!spaf
From: spaf@gatech.UUCP (Gene Spafford)
Newsgroups: net.lang
Subject: Re: Self-modifying code - (nf)
Message-ID: <3141@gatech.UUCP>
Date: Sun, 8-Jan-84 12:24:16 EST
Article-I.D.: gatech.3141
Posted: Sun Jan  8 12:24:16 1984
Date-Received: Mon, 9-Jan-84 07:01:36 EST
References: <692@inmet.UUCP>
Organization: Georgia Tech School of ICS, Atlanta
Lines: 36

This concerns the testing method whereby the code is modified with trap
instructions.

Richard DeMillo (a faculty member here at Georgia Tech) and some of his
students did some work in the field of "program mutation" a while
back.  This scheme is akin to code modification with traps.  Basically,
what program mutatiion does is "mutate" the source code and then run a
test suite on the mutations.  "Viable" mutations are then saved for
further analysis.  Mutations are things like changing loop boundaries,
re-ordering certain statements, and putting halt statements in sections
of code that might never be reached.

When a program had undergone a suite of mutation tests, the resulting
program which behaved correctly could either be further tested, or
compared by a human being.  Some mutations might be found to be
"benign" and of no further interest, but some mutations might point out
flaws in the original coding or sections of code which were never
executed.

As I understand it, this idea was actually applied to large sets of
Cobol and Fortran programs, and may still be in use for some
applications (I wasn't involved with the project in any way, so I don't
know).  Since it goes beyond simply replacing each instruction with a
"trap," I thought it might be of interest.  The bibliographies to the
tech reports might also be of interest.  If anybody is interested in
more information, I suggest you write to:
	Richard DeMillo
	School of Information and Computer Science
	Georgia Institute of Technology
	Atlanta, GA 30332
and ask about his program mutation project.
-- 
Off the Wall of Gene Spafford
The Clouds Project, School of ICS, Georgia Tech, Atlanta GA 30332
CSNet:	Spaf @ GATech		ARPA:	Spaf.GATech @ CSNet-Relay
uucp:	...!{akgua,allegra,rlgvax,sb1,unmvax,ulysses,ut-sally}!gatech!spaf