Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!sdd.hp.com!spool.mu.edu!news.cs.indiana.edu!purdue!bouma
From: bouma@cs.purdue.EDU (William J. Bouma)
Newsgroups: comp.lang.forth
Subject: Re: One and One doesn't equal two It equals 1 - The Who
Message-ID: <14079@medusa.cs.purdue.edu>
Date: 27 Mar 91 20:38:29 GMT
Article-I.D.: medusa.14079
References: <9103271403.AA16095@ucbvax.Berkeley.EDU>
Sender: news@cs.purdue.EDU
Organization: Department of Computer Science, Purdue University
Lines: 38

In article <9103271403.AA16095@ucbvax.Berkeley.EDU> "Frank C. Earl" <S47852EF%ETSUACAD.BITNET@SCFVM.GSFC.NASA.GOV> writes:
>On Mon, 25 Mar 91 10:43:42 GMT <zmleb@SCFVM.GSFC.NASA.GOV> said:
>>
>>Sorry, but any Scheme or Lisp I know of will complain that '1' is not a symbol
>
>Indeed...    But, Forth will let you do it...   (I just LOOOVE that!   ;)

   The Forth philosophy seems to be that everything read in is a WORD.
   Upon reading "1" and not finding an entry in the dictionary, the 
   'default` action of "1" is performed.  Namely, to push the value 1
   on the stack.  This is consistent with Forth's refusal to acknowledge
   there are varied types of data.  The interpreter refuses to read
   anything other than WORDs.

   Lisp simply has a different philosophy.  Lisp is typed.  The reader
   returns the value 1 when it sees "1", rather than the symbol who's
   name is "1".  The reader returns a vector when it sees #() and a list
   when it sees '().  To have the reader return the symbol who's name is
   "1", the syntax (oh, no, that swear word again) is |1|.  In lisp, (I
   am really talking only CL here, don't know how Scheme does it), I
   could thus say (defun |1| () 2) or (setq |1| 2) without error.
   
   Notice Forth would need a special syntax to distinguish numbers and
   names of numbers if its search order were changed.  ie. Check if the
   string is a number first, if not, search in the dictionary.

   I have accidentally redefined a constant in FORTRAN before.  One did
   (do? depends on the implementation I guess) this by calling a subroutine
   with a constant argument.  Since FORTRAN passes arguments by address...
   This is a tough bug to find!  However, I find it hard to believe anyone
   could accidentally do this in Forth.  If you did misstype the name to a
   colon definition, it would be an easy bug to spot.

   You can love or hate that Forth allows you to do this.  It is novel but
   potentially dangerous.  What I dislike is the philosophy of Forth that
   makes it possible to allow you to do it. 8^)  But you know that already.
-- 
Bill <bouma@cs.purdue.edu>