Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!uunet!seismo!husc6!bloom-beacon!tdavis
From: tdavis@athena.mit.edu (Timothy L. Davis)
Newsgroups: comp.lang.c
Subject: Re: goto's
Message-ID: <1217@bloom-beacon.MIT.EDU>
Date: Wed, 29-Jul-87 02:53:25 EDT
Article-I.D.: bloom-be.1217
Posted: Wed Jul 29 02:53:25 1987
Date-Received: Fri, 31-Jul-87 01:35:57 EDT
References: <3289@bigburd.PRC.Unisys.COM> <7571@beta.UUCP> <765@haddock.ISC.COM> <7687@mimsy.UUCP> <465@bms-at.UUCP>
Sender: daemon@bloom-beacon.MIT.EDU
Reply-To: tdavis@ATHENA.MIT.EDU (Timothy L. Davis)
Organization: Massachusetts Institute of Technology
Lines: 22

In article <465@bms-at.UUCP> stuart@bms-at.UUCP (Stuart D. Gathman) writes:
>
>Knuth has an excellent method for converting any program, no matter how
>convoluted, to perfectly structured code with no goto's.
>...
>    IF P = 1 THEN
>      (* DO STEP 1 *)
>      P := (* STEP TO PERFORM AFTER 1 *)
>    IF P = 2 THEN
>      (* DO STEP 2 *)
>      P := (* STEP TO PERFORM AFTER 2 *)
>...

I am just finishing a modification of a program which used this precise
system for "computed" gotos.  What a mess!  It was nearly impossible to read 
clearly. The little assignment statements, which were in effect gotos, were 
MUCH harder to follow than goto's themselves.  If you are going to write 
spaghetti code, you might as well use the spaghetti construct!  

Tim Davis
tdavis@ATHENA.MIT.EDU
...mit-eddie!mit-athena!tdavis