Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.1 6/24/83 SMI; site sun.uucp
Path: utzoo!watmath!clyde!burl!ulysses!harpo!decvax!decwrl!sun!gnu
From: gnu@sun.uucp (John Gilmore)
Newsgroups: net.lang.apl
Subject: Re: iteration in the expression sub-language of APL
Message-ID: <1089@sun.uucp>
Date: Wed, 16-May-84 01:25:38 EDT
Article-I.D.: sun.1089
Posted: Wed May 16 01:25:38 1984
Date-Received: Thu, 17-May-84 05:09:33 EDT
References: <8059@cornell.UUCP>
Organization: Sun Microsystems, Inc.
Lines: 15

Several APL systems have extensions like what you propose.  Systems from IBM,
STSC, and IPSA all include an extension to compression called "replicate";
each left argument element says how many copies of the right argument element
should appear in the result.

Also, for multiple application of functions, the "power" operator is used.
I don't recall exactly how it works for dyadics, but it's easy for monadics.
It just applies the function N times where N and the function are the arguments
to the operator.  There are degenerate cases where N is missing (apply the
function until the result stops changing, as in calculus) or where N is -1
(give the inverse function).

I'd suggest getting a copy of the IBM APL2 Language Manual to get some idea
of what is going on in the modern APL world.  You might also check some of
the "nested arrays and operators" papers from the last 5 or 6 APL conferences.