Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!usc!elroy.jpl.nasa.gov!ncar!hsdndev!cmcl2!adm!smoke!gwyn
From: gwyn@smoke.brl.mil (Doug Gwyn)
Newsgroups: comp.lang.c
Subject: Re: Small introspective program
Message-ID: <15491@smoke.brl.mil>
Date: 16 Mar 91 22:40:50 GMT
References: <4963@goanna.cs.rmit.oz.au> <217@hermix.UUCP> <1991Mar16.163213.9794@bingvaxu.cc.binghamton.edu>
Organization: U.S. Army Ballistic Research Laboratory, APG, MD.
Lines: 19

In article <1991Mar16.163213.9794@bingvaxu.cc.binghamton.edu> kym@bingvaxu.cc.binghamton.edu (R. Kym Horsell) writes:
>The following is a (probabalistic) example. It relies on fixpoint
>theory to guarantee finding the appropriate parameters (in fact the
>search for any self-replicating automaton can use this theory) and 
>terminates in several iterations.

I think it was Ken Thompson who tried the experiment of taking the
diagnostic output from a UNIX C compiler and feeding it back to
another invokation of the compiler as source code:

	echo foo bar > X	# initial invalid "program"
	while	:
	do	cp X x.c
		cc x.c > X 2>&1 || :
		if cmp -s X x.c
		then	cat X
			exit
		fi
	done