Path: utzoo!utgpu!jarvis.csri.toronto.edu!clyde.concordia.ca!uunet!microsoft!jimad
From: jimad@microsoft.UUCP (JAMES ADCOCK)
Newsgroups: comp.lang.c++
Subject: Re: const enums -- Lawyers opinions??
Keywords: const, enum
Message-ID: <10215@microsoft.UUCP>
Date: 10 Jan 90 17:36:40 GMT
References: <963@rna.UUCP> <10197@microsoft.UUCP> <9925@ardent.UUCP>
Reply-To: jimad@microsoft.UUCP (JAMES ADCOCK)
Organization: Microsoft Corp., Redmond WA
Lines: 24

>The question of what an enum type "really is",  has bedeveled
>the concept ever since it was introduced into C.  Over the
>years a lot has been said, but I believe a consensus finally
>emerged and both ANSI C and the most recent C++ reference manual agree 
>that they are integral types.  
>
>This means that its perfectly sensible to assign arbitrary integer
>values to variables of enum type. 
>
>Jerry Schwarz

I would claim on the contrary that the decision by both ANSI-C and C++
to consider enums integral types is a pragmatic decision stating that
strict type checking on enums is more of a pain than a help for 
programmers.  I would claim that if one creates an enum for the days
of the week and then store the integer 10000 in an instance of that
enum, then you've made a programming mistake -- even if it is permissable
according to the language definition.  There are a lot of areas in C++
like that -- where the language for historical or pragmatic reasons allows
programmers to do things that are generally very bad programming practice.

I believe it is important that programs "scan" right to human readers --
if a variable claims to be of a certain enum, it should only hold legal
values of that enum.  To do otherwise is a bug.  IMHO.