Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/18/84; site brl-tgr.ARPA
Path: utzoo!watmath!clyde!burl!ulysses!allegra!mit-eddie!godot!harvard!seismo!brl-tgr!ron
From: ron@brl-tgr.ARPA (Ron Natalie <ron>)
Newsgroups: net.lang.c
Subject: Re: Re:  Re:  hashing external names & other goodies.
Message-ID: <7187@brl-tgr.ARPA>
Date: Thu, 10-Jan-85 15:43:37 EST
Article-I.D.: brl-tgr.7187
Posted: Thu Jan 10 15:43:37 1985
Date-Received: Sat, 12-Jan-85 07:39:48 EST
References: <7140@brl-tgr.ARPA>
Organization: Ballistic Research Lab
Lines: 22


> |And "goto <label>" is even better, you don't have to change the language
> |at all.
> |
> |-Ron
>
> Ron, perhaps you missed the point. The major objection to GOTO's is that they
> can lead to spaghetti code, but "break <label>" cannot do so, as it allows
> only a clearly limited type of downward escape, while avoiding the pitfalls
> of having to count nesting levels.
> 
> David   sde@mitre-bedford

I do understand the point, but BREAK-TO is not structred anymore than
any sort of gotos.  If you just say, it's OK to user goto's rather only
in these well-defined casees, you can get by WITHOUT MAKING GRATUITOUS
AND UNNECESSARY CHANGES TO THE LANGUAGE.  BREAK-TO is a total unnecessary
change.  You're not going to be able to get rid of goto (one of the nice
things about C is that it allows careful programmers to break the structure
and typing rules), adding break-to is redundant.

-Ron