Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.1 6/24/83; site bu-cs.UUCP
Path: utzoo!watmath!clyde!cbosgd!ihnp4!mhuxn!mhuxm!sftig!sftri!sfmag!eagle!ulysses!unc!mcnc!decvax!genrad!panda!talcott!harvard!bu-cs!hen
From: hen@bu-cs.UUCP (Bill Henneman)
Newsgroups: net.ai
Subject: Re: Thus spake the DoD...
Message-ID: <215@bu-cs.UUCP>
Date: Fri, 8-Mar-85 12:40:25 EST
Article-I.D.: bu-cs.215
Posted: Fri Mar  8 12:40:25 1985
Date-Received: Mon, 11-Mar-85 04:13:53 EST
References: <313@wdl1.UUCP>, <215@rtech.ARPA>
Organization: Boston Univ Comp. Sci.
Lines: 15

I think it would be a *rare* piece of Lisp code that could be run
without most of the functionality of LISP at run time: all of the
implimentations I am familiar with ensure that anything that builds list
structure funnel through a call to CONS.  CONS uses an implicit
assumption of a LISP garbage collector present when the CONS can't find
space.  But once you have the memory management routines (and all that
they imply) resident at run time, you have come so close to having a
LISP, you might as well throw in the rest (READ is the only other hairy
thing to write).

I have been faced with this issue in conjunction with an automatic
code-porting project I have been working on this year; my current state
is that I believe one can write systems to take code in any compiled
language and produce equivalent LISP code, but that converting in the
opposite direction is not economically feasible.