Newsgroups: comp.lang.dylan
Path: cantaloupe.srv.cs.cmu.edu!rochester!udel!gatech!news.sprintlink.net!mv!condes.mv.com!user
From: gmcgath@condes.mv.com (Gary McGath)
Subject: Re: Why no null?
Message-ID: <gmcgath-2107950726420001@condes.mv.com>
Nntp-Posting-Host: condes.mv.com
Sender: usenet@mv.mv.com (System Administrator)
Organization: Conceptual Design
Date: Fri, 21 Jul 1995 11:26:42 GMT
References: <Johan.Dahl-1907951710350001@venus.ling.lu.se>
Lines: 19

In article <Johan.Dahl-1907951710350001@venus.ling.lu.se>,
Johan.Dahl@ling.lu.se (Johan Dahl) wrote:

> I have wondered why the designers of Dylan choosed to not include some
> kind of null or nil constant. Dylan programs seems to use either a type
> union including the singelton #f as representing no object or making two
> subclasses one which defines the object and one which defines the "no
> object".

While I'm by no means a Dylan expert, I'm also puzzled by the absence. The
#f mechanism to indicate an absence is clumsy at best. I suppose the
absence of a null object simplifies error checking; but given that a Dylan
implementation supports checking for uninitialized references, could it be
that much more work to support null references?

-- 
          Gary McGath
          gmcgath@condes.mv.com
PGP Fingerprint: 3E B3 62 C8 F8 9E E9 3A  67 E7 71 99 71 BD FA 29
