Newsgroups: comp.lang.scheme
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!howland.reston.ans.net!news.sprintlink.net!siemens!princeton!news.princeton.edu!blume
From: blume@atomic.cs.princeton.edu (Matthias Blume)
Subject: Re: The UObj Puzzle -- my (flawed) solution [as promised]
In-Reply-To: bh@anarres.CS.Berkeley.EDU's message of 7 Feb 1995 00:47:05 GMT
Message-ID: <BLUME.95Feb7100401@atomic.cs.princeton.edu>
Originator: news@hedgehog.Princeton.EDU
Sender: news@Princeton.EDU (USENET News System)
Nntp-Posting-Host: atomic.cs.princeton.edu
Organization: Princeton University
References: <9502061918.aa18888@mc.lcs.mit.edu> <3h6fu9$ebp@agate.berkeley.edu>
Date: Tue, 7 Feb 1995 15:04:01 GMT
Lines: 28

In article <3h6fu9$ebp@agate.berkeley.edu> bh@anarres.CS.Berkeley.EDU (Brian Harvey) writes:

   ziggy@martigny.ai.mit.EDU (Michael R. Blair) writes:
   >Alas, even this has a subtle bug...

   Geez.  Okay, if even ziggy can't get it right without weeks of thinking
   about it, doesn't this indicate something is screwy about the alleged
   feature that makes this so hard?

Probably.

   If it were me, I'd say that every invocation of lambda should be required
   to make an EQ?-unique object, but that if a compiler can prove that in fact
   nobody calls EQ? on some batch of procedures, then it's allowed to be clever.

And if it was me, I'd say that every invocation of EQ? on an object
constructed by LAMBDA should either be an error, or, if not that,
should at least be left unspecified.  But you must have guessed that
already... :)

   (And I suppose you'd be allowed to have some nonstandard pragma thingo as
   for integrable and the like.)

Some nonstandard *WHAT*?!  The word `nonstandard' alone makes me
shudder.

--
-Matthias
