Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!helios!willis
From: willis@cs.tamu.edu (Willis Marti)
Newsgroups: comp.lang.misc
Subject: Re: The Forbiden
Summary: restrictions are necessary
Message-ID: <7662@helios.TAMU.EDU>
Date: 23 Aug 90 19:40:30 GMT
References: <24279@megaron.cs.arizona.edu>
Sender: usenet@helios.TAMU.EDU
Organization: Computer Science Department, Texas A&M University
Lines: 23

In article <24279@megaron.cs.arizona.edu> gudeman@cs.arizona.edu (David Gudeman) writes:

[lots of other stuff deleted]

>impossible in most non-recursive programs.  The only programs where
>non-recursion cannot be proven automatically are those which have a
>recursive call in a section of code that will never be executed, but
>where the compiler cannot prove that the code will never be executed.
>Since such programs are almost guaranteed to be not what the
>programmer intended, I feel confident in saying that non-recursion can
>be automatically proven in correct non-recursive programs.

Not true.  In C, with pointers to functions; in assembler, with indirect
jumps; in FORTRAN, with computed GOTOs -- there are MANY times a compiler
can't "prove" the execution path of a program without running it.  And some
of those paths may use recursion, some may never use it.  I feel you really
have the choice of the language designer making some assumptions 
(i.e., forbidding some coding practices) or making a language that's either
extremely verbose (can you  say PL/1 or Ada? 8-) or provides zero checking...
Programmers should choose an appropriate tool & not complain that their one
tool doesn't have feature XYZ.

Willis Marti