Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.1 6/24/83 (MC840302); site zuring.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!allegra!mit-eddie!think!harvard!seismo!mcvax!zuring!dik
From: dik@zuring.UUCP
Newsgroups: net.lang
Subject: Re: Recursion
Message-ID: <240@zuring.UUCP>
Date: Sat, 31-Aug-85 00:44:55 EDT
Article-I.D.: zuring.240
Posted: Sat Aug 31 00:44:55 1985
Date-Received: Sun, 1-Sep-85 05:55:23 EDT
References: <909@oddjob.UUCP> <163@ho95e.UUCP> <367@ttrdc.UUCP> <1612@watdcsu.UUCP> <371@ttrdc.UUCP> <605@bu-cs.UUCP> <624@mmintl.UUCP>
Reply-To: dik@zuring.UUCP (Dik T. Winter)
Organization: CWI, Amsterdam
Lines: 16
Apparently-To: rnews@mcvax.LOCAL

In article <624@mmintl.UUCP> franka@mmintl.UUCP (Frank Adams) writes:
>
>Concerning the recent discussion of FORTRAN and recursion: I think recursive
>routines should have to be declared as such, as in (gasp) PL/I.  Not because
>of efficiencies gained by knowing when a routine is recursive, but as an aid
>in writing correct code.  ...

Of course, but in than case you have a new problem.  The system should check
whether the routine you declared as non-recursive is non-recursive indeed.
(What about routine 1 calling routine 2; which in turn calls routine 1.
I have had uses for this in a non-recursive way.)
Moral: a system (or a language) should never require things it is not
able to ckeck (like Fortran).
-- 
dik t. winter, cwi, amsterdam, nederland
UUCP: {seismo|decvax|philabs}!mcvax!dik