Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!allegra!mit-eddie!godot!ima!inmet!bobo
From: bobo@inmet.UUCP
Newsgroups: net.lang
Subject: Re: Nil dereferencing protection
Message-ID: <1934@inmet.UUCP>
Date: Thu, 31-Jan-85 03:42:23 EST
Article-I.D.: inmet.1934
Posted: Thu Jan 31 03:42:23 1985
Date-Received: Sat, 2-Feb-85 09:49:12 EST
Lines: 13
Nf-ID: #R:harvard:-33800:inmet:4700021:000:669
Nf-From: inmet!bobo    Jan 29 15:37:00 1985

> > > For instance, it is possible to statically guarantee that nil
> > > pointers are never dereferenced  --Pemberton
> > Ada supports this...  --Steinman

> Unfortunately, this is not and cannot be true.  (consider "a := if
> program P halts, then nil else new foo; deref(a)")

Excuse me for putting words into Mr. Pemberton's terminal but I
think that what he was trying to say was NOT that you can guarantee
that a nil dereference WILL occur, but that it WILL NOT occur.
In the above example a compiler can warn you that a nil dereference
MAY occur. In many situations this information can help you to
discover bugs. In other situations you may just say "so what."