Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.1 6/24/83; site ho95b.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!mhuxr!mhuxt!houxm!ho95b!wcs
From: wcs@ho95b.UUCP (Bill Stewart)
Newsgroups: net.lang.c
Subject: Re: "Vectorizing C compiler for the Cray" (Call-by-reference)
Message-ID: <388@ho95b.UUCP>
Date: Tue, 23-Apr-85 20:50:20 EST
Article-I.D.: ho95b.388
Posted: Tue Apr 23 20:50:20 1985
Date-Received: Wed, 24-Apr-85 05:11:56 EST
References: <486@lll-crg.ARPA> <515@lll-crg.ARPA> <> <330@cubsvax.UUCP>
Organization: AT&T Bell Labs, Holmdel NJ
Lines: 27

I suspect that the motivation for a call-by-address C is to have
something more useful than FORTRAN-66, but that won't confuse fortran
programmers.  A couple of solutions, more reasonable than 
		brain-damaging C -

- Fortran-77	The language assumes call-by-reference, and could
		easily enough tolerate having recursion added (as UNIX
		f77 does.)  If all you really want are control
		structures and recursion, this is a good place to start,
		and it will keep old programs working.  I suppose yu
		could also add pointers - the resulting language would
		resemble fortran more than call-by-ref resembles C.

- PL/I		Well, some reasonable-sized subset of PL/I.  The
		language has pointers, of sorts.  There are lots of
		six-foot-thick manuals for the language, and it
		has a few features useful for scientific programming
		that C doesn't have, specifically arrays with variable
		dimensions (you can declare A(N,N), where N is a
		parameter passed to a subroutine).

- APL		an inherently vector-oriented language should be easy
		to adapt to a vector-processor, and you it's a nice
		language for expressing mathematics, and for debugging
		on your "slow" computer before you turn it loose on the
		real data.
				Bill Stewart