Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.1 6/24/83; site burdvax.UUCP
Path: utzoo!linus!vaxine!wjh12!genrad!decvax!mcnc!akgua!psuvax1!burdvax!csgt
From: csgt@burdvax.UUCP (Steve Bartels)
Newsgroups: net.lang
Subject: Re: APL Myths
Message-ID: <1588@burdvax.UUCP>
Date: Wed, 27-Jun-84 00:27:45 EDT
Article-I.D.: burdvax.1588
Posted: Wed Jun 27 00:27:45 1984
Date-Received: Sat, 30-Jun-84 06:56:34 EDT
Organization: System Development Corporation, Paoli PA
Lines: 36

A possible reason why people find APL hard to read is that they compare
the time to understand 1 token in APL vs 1 token in other languages,
without realizing that each APL token is about equivalent to 15 or 20
tokens in Fortran, Algol, etc.  If they compared the time to understand
a program in APL vs a program in another language, APL will probably 
win.

With structure programming, I usually end up with lots of little
procedures that do higher-level operations on higer-level data types
than the language directly supports.  It is only recently that I have
realized that I have actually been defining another language, one that
is more suitable for my particular application needs.  The result is
that anyone looking at my code (or at anyone's SP code) has to know
two languages, this higher-level language and the base language.  And
each program will have a slightly different higher-level language!

APL avoids most of this by having the base language powerful enough
to support very high-level data types with only a few characters.
You don't need to define the little functions since it is easier to
enter what you need right on the line, and avoids any conflicts
between the two languages.

I know that there is alot more to SP than what I have been talking 
about.  However, this is something I noticed I have been doing in
ALGOL and PASCAL alot, and not in APL.  Reading other peoples programs
in PASCAL is a pain because I have to look up all these little
procedures to get their exact definition.  I don't have to look
up anything in APL.

Stephen R. Bartels
(APL programmer for 12 years, implementator for 6)
...!burdvax!csgt
Subject: Re: APL myths
Newsgroups: net.lang