Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!uunet!munnari.oz.au!metro!usage.csd.unsw.oz.au!syacus!ian
From: ian@syacus.acus.oz.au (Ian Joyner)
Newsgroups: comp.lang.pascal
Subject: Re: Goto's are okay
Keywords: goto
Message-ID: <1991Feb3.231920.8462@syacus.acus.oz.au>
Date: 3 Feb 91 23:19:20 GMT
References: <2427@bnlux0.bnl.gov> <1991Jan27.074528.23487@tygra.UUCP> <1991Jan27.185044.19856@uwasa.fi> <1991Jan28.102509.5587@tygra.UUCP>
Organization: ACUS Australian Centre for Unisys Software, Sydney
Lines: 21

There is yet another consideration in the goto debate. This is
concurrency. The use of gotos make it difficult for a compiler to generate
concurrent code. While this may not be a problem at the moment, I guess
machines coming out in the next ten years will be inherently concurrent.

This is also the argument against global variables. You will have to
synchronise updates to globals, and if many processes are trying to
update the same set of variables, this will result in a bottleneck,
or worse still be error prone.

Avoiding globals is easy. If you find you have the urge to put in some
globals, chances are that you have found another class. If not they should
all be placed in an "application" class. Of course it is impossible to
avoid globals where the language isn't a suitable tool. I don't find
languages that don't have gotos or globals restrictive. It will probably
be common in the future, and noone will argue, because such languages
make the new generation of parallel and concurrent systems possible.
-- 
Ian Joyner                                     ACSNet: ian@syacus.oz
ACUS (Australian Centre for Unisys Software)   DNS:  ian@syacus.oz.au
Tel 61-2-390 1328      Fax 61-2-390 1391       UUCP: ...uunet!munnari!syacus.oz