Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!seismo!husc6!mit-eddie!genrad!decvax!ucbvax!cbatt!cwruecmp!arun
From: arun@cwruecmp.UUCP (Arun Lakhotia)
Newsgroups: comp.lang.prolog
Subject: Self reproducing prolog program
Message-ID: <2073@cwruecmp.UUCP>
Date: Sun, 3-May-87 23:33:31 EDT
Article-I.D.: cwruecmp.2073
Posted: Sun May  3 23:33:31 1987
Date-Received: Tue, 5-May-87 00:46:24 EDT
Reply-To: arun@cwruecmp.UUCP (Arun Lakhotia)
Distribution: world
Organization: CWRU Dept. of Computer Engineering, Cleveland, Ohio
Lines: 30

> clone :-
>        listing(clone).

May be i am being too picky.

The above program would produce all clauses having
head with functor 'clone', which could be more than
re-producing if the program database have more clone/?
clauses.

I am not sure whether that violates the notion of 
self-reproducibility, ie the offspring subsumes the
parent, and is not exactly equal but more :-) than equal.

Just an observation here ..

Another important requirement for a s/r program normally 
stated is that 'it should not make use of any knowledge
of the implementation of the language, or the machine
its implemented on'. Thus the s/r program should 
reproduce itself on any reasonable implementation
of the language in question.

With the given restriction, wouldn't it be necessary that any
s/r program be free of Variables. As most implementations
rename variables and show some _XXX kind of number on
listing or write, it would be kind of difficult, (if not
impossible) to write a generic s/r program.
=
arun