Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.1 6/24/83; site mmintl.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!allegra!mit-eddie!genrad!panda!talcott!harvard!cmcl2!philabs!pwa-b!mmintl!franka
From: franka@mmintl.UUCP (Frank Adams)
Newsgroups: net.lang
Subject: Re: Recursion
Message-ID: <661@mmintl.UUCP>
Date: Mon, 16-Sep-85 19:42:16 EDT
Article-I.D.: mmintl.661
Posted: Mon Sep 16 19:42:16 1985
Date-Received: Fri, 20-Sep-85 06:33:33 EDT
References: <712@gitpyr.UUCP> <250@mot.UUCP> <241@zuring.UUCP>
Reply-To: franka@mmintl.UUCP (Frank Adams)
Organization: Multimate International, E. Hartford, CT
Lines: 15

In article <241@zuring.UUCP> dik@zuring.UUCP (Dik T. Winter) writes:
>(about A calling B calling A, and how to detect this)...
>
>No, the 1's indicate the routines that *might* be recursive, they
>need not be.  (What in the situation that if A calls B, B will never
>call A, vv.)

I have encountered this situation.  My feeling is that usually, it is the
result of sloppy design.  There are probably cases where it is legitimate,
however.

I would have no problem with labelling the routines A and B 'recursive'
in this case.  It is still true that if A calls B and B calls A, unless
it was recognized in the design that this would happen, it is nearly
certain that this is a bug.