Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.3 4.3bsd-beta 6/6/85; site gtss.UUCP
Path: utzoo!linus!gatech!gtss!chas
From: chas@gtss.UUCP (Charles Cleveland)
Newsgroups: net.jokes
Subject: Re: Computer Bloopers
Message-ID: <131@gtss.UUCP>
Date: Mon, 5-Aug-85 10:00:25 EDT
Article-I.D.: gtss.131
Posted: Mon Aug  5 10:00:25 1985
Date-Received: Tue, 6-Aug-85 03:50:13 EDT
References: <3425@dartvax.UUCP>
Reply-To: chas@gtss.UUCP (Charles Cleveland)
Distribution: net
Organization: Georgia Tech Surface Studies
Lines: 24

In article <3425@dartvax.UUCP> merchant@dartvax.UUCP (Peter Merchant) writes:
>Well, the one I heard that I always liked is one of the old FORTRAN compilers
>that would allow you to change the value of constants if you passed one to
>a subroutine (or whatever FORTRAN calls them) because EVERYTHING was passed
>by reference.

These compilers aren't all old.  Just recently, we had some execution problems
on a Cyber mainframe that kept complaining that unit 13 wasn't open, so we
couldn't write on it.  Of course, we were writing on unit 6, 6 had been passed
to a subroutine (having nothing to do with I/O of course) as one of its
arguments used for both input and output, etc. etc.

Naturally, this isn't a problem to track down any more -- its not the kind of
thing you forget.

It occurs to me that the thorough application of this principle could produce
one hell of a contender in an obfuscated FORTRAN contest.

-- 
Charles Cleveland			 Georgia Tech Surface Studies
Georgia Tech School of Physics
Atlanta, GA 30332
...!{akgua,allegra,amd,hplabs,ihnp4,masscomp,ut-ngp}!gatech!gtss!chas
...!{rlgvax,sb1,uf-cgrl,unmvax,ut-sally}!gatech!gtss!chas