Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!rutgers!apple!bloom-beacon!SESAME.STANFORD.EDU!mkatz
From: mkatz@SESAME.STANFORD.EDU (Morris Katz)
Newsgroups: comp.lang.scheme
Subject: Scheme Digest #125
Message-ID: <8905301538.AA06079@sesame.Stanford.EDU>
Date: 30 May 89 15:38:17 GMT
References: <wYTTsRC00Vsn8KomQr@andrew.cmu.edu>
Sender: daemon@bloom-beacon.MIT.EDU
Organization: The Internet
Lines: 26


   Date: 27 May 89 01:47:09 GMT
   From: Robert Steven Glickstein <bobg+@andrew.cmu.edu>

   > Excerpts from ext.nn.comp.lang.scheme: 25-May-89 Scheme Digest #125
   > Jonathan Shapiro@POLYA.S (631)

   > Very well.  I hereby specify an object whose name is

   > 	#UNSPECIFIED

   I second the motion.  I am currently designing ELSIE (the "Embedded Lisp
   [Scheme] Interpreter, `ELSIE'"), and I'd rather not have to make
   arbitrary decisions about what SET! returns (for instance).  (Actually,
   set! is a bad example, since I believe it should return the new value,
   which I guess I'm proposing for R4RS, but you get the idea.)

   -Bob Glickstein

Since the return value is left unspecified, you don't have to make an arbitrary
decision.  Simply created #UNSPECIFIED in your implementations and return it.
This is completely consistent with the current standard.  Just don't force
#UNSPECIFIED on the rest of our implementations. 

					Morry Katz
					katz@polya.stanford.edu