Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/18/84; site ucsfcgl.UUCP
Path: utzoo!watmath!clyde!cbosgd!ihnp4!mhuxn!mhuxb!mhuxr!ulysses!allegra!mit-eddie!godot!harvard!talcott!panda!genrad!decvax!ucbvax!ucsfcgl!arnold
From: arnold@ucsfcgl.UUCP (Ken Arnold%CGL)
Newsgroups: net.lang
Subject: Re: Turing the first?
Message-ID: <436@ucsfcgl.UUCP>
Date: Sat, 2-Feb-85 23:45:25 EST
Article-I.D.: ucsfcgl.436
Posted: Sat Feb  2 23:45:25 1985
Date-Received: Wed, 6-Feb-85 04:41:46 EST
References: <8900018@uiucdcsb.UUCP> <18218@lanl.ARPA> <428@ucsfcgl.UUCP> <870@pucc-i>
Reply-To: arnold@ucsfcgl.UUCP (PUT YOUR NAME HERE)
Organization: UCSF Computer Graphics Lab
Lines: 46
Summary: 

In article <870@pucc-i> ags@pucc-i (Dave Seaman) writes:
>Turing was not the first to think of self-modifying code.  Charles Babbage,
>the inventor of the Analytical Engine, wrote about this concept in the
>1840s.  His term for it was "the Engine eating its own tail."
>
>Babbage's colleague Ada Augusta, the Countess of Lovelace (for whom the
>Ada programming language is named) even anticipated the possibility of
>machine translation of symbolic programs (i.e. compilers).
>-- 
>Dave Seaman			 ..!pur-ee!pucc-i:ags

I admit not to having read Babbage's original paper on this, but my
understanding was that he used the phrase "eating its own tail" to
refer to conditional branches based on its own calculated values.  I
was talking about code which modified itself.  Below is a quote from a
1950's history of computing (I don't have the reference here at home,
but will be happy to provide it upon request) which describes Babbage's
concept:

	Babbage clearly understood the restrictions imposed by the
	inability of a machine to make decisions for itself.  He was
	able to take the next step and to suggest how to endow a
	machine with the minimum amount if "intelligence" which it
	needs.  As he expressed it, he made the machine "bite its own
	tail". ...

	If a machine is to perform the functions of a human computer
	[used here in the old sense, i.e., "one who computes" --KA], it
	must possess --

	(a) An arithmetic unit ...  (b) A memory ...  (c) A built-in
	power of judgement, which will enable the machine
	    to choose, according to prescribed criteria, the course
	    which the computation has to take.  (d) An input-ouput
	mechanism ...

This kind of decision is solely conditional branching; it is choosing
of alternatives.  Turing's concept was of the creation and modification
of programs themselves according, thus it is "self-modifying", as
distinct from Babbage's "self-controlling".
-- 

		Ken Arnold
=================================================================
Of COURSE we can implement your algorithm.  We've got this Turing
machine emulator...