Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!samsung!zaphod.mps.ohio-state.edu!rpi!uupsi!grebyn!dlawson
From: dlawson@grebyn.com (Drew Lawson)
Newsgroups: comp.theory
Subject: Re: Procedures with multiple out parameters: why not?
Message-ID: <21458@grebyn.com>
Date: 16 Aug 90 14:36:08 GMT
References: <1990Aug15.231838.5664@zaphod.mps.ohio-state.edu>
Reply-To: dlawson@grebyn.UUCP (Drew Lawson)
Organization: Grebyn Timesharing, Vienna, VA, USA
Lines: 28

In article <1990Aug15.231838.5664@zaphod.mps.ohio-state.edu> vidynath@function.mps.ohio-state.edu (Vidhyarath K Rao) writes:
>
>
>While there is a lot of well-known [i.e. I have heard of them :-^] articles
>on the goto statement, I have never seen heard of the theoreticians' case
>against having more than one out parameter. None of the reasons I could
>think of relates to theory. Can somebody out there give some pointers?
>

I don't have any references to this, but the attitude I have beed
exposed to which agrees with it is that procedures/functions should
not both return a value and alter parameters.  The problems can be similar
to the goto problems: it is less obvious what the flow is (in this case
the data flow rather than the control flow).

I don't cling too tightly to this rule.  Being a C programmer, I make a
lot of use of procedures with variable parameters which return a logical
value indicating whether the operation was completed correctly.

I don't know of any strictly theoretical backing for Dijkstra's assertion,
just pragmatic issues.  But then, a large part of the problems with gotos
is pragmatics.

-- 
+--------------------------------------------------------------------+
| Is life an illusion?                          | Drew Lawson        |
| Or does it just seem that way?                | dlawson@grebyn.com |
+--------------------------------------------------------------------+