Newsgroups: comp.lang.apl
Path: watmath!ljdickey
From: ljdickey@math.uwaterloo.ca (L.J. Dickey)
Subject: Re: ascii version of apl
Message-ID: <C7nJ68.1Dv@math.uwaterloo.ca>
Sender: news@math.uwaterloo.ca (News Owner)
Organization: University of Waterloo
References: <m47q3B1w165w@ersys.edmonton.ab.ca>
Date: Wed, 26 May 1993 20:49:19 GMT

In article <m47q3B1w165w@ersys.edmonton.ab.ca> ab@ersys.edmonton.ab.ca (Allan Brockman) writes:

> ... i would like to know why i have not seen 
> someone sugest using j as a ascii representation of APL?


J and APL are too far apart.  The easiest example I can mention has to
do with replicate and reduce.  You can not tell from the APL code,
(unless you parse it, and are about to execute it),
whether the slash  "/"  means replicate or reduce.

First, compare these two APL examples:

		     1 2 3 / 7 8 9			.comment: replicate
		7 8 8 9 9 9
		     + / 7 8 9				.comment: reduce
		24

Second, take a look at this trouble spot: if you encounter the line
		     F / DATA
by itself, you can not tell whether F is supposed to represent some
numeric data, in which case the "/" means replicate and should be
translated to a "#" in J, or whether F is a user defined function, in
which case ( F / ) is a derived function.

You might argue that one should be able test the name class of F, and
then take appropriate action.  Sometimes this strategy works, but it
fails if the line in question is buried inside a function or in a
character string that gets executed at a later time in some other
environment in which the name class of a new "F" is different.

I like the fact that J has separated the two different meanings of
the APL "/".  It is unfortunate that this ambiguity exists in APL.

I like the transliteration scheme proposed by Johann Mitloehner of
Vienna and Sam Sirlin of Pasadena.  These guys have done good work.

Lee Dickey


-- 
Prof. Leroy J. Dickey, Faculty of Mathematics, U of Waterloo, Canada  N2L 3G1
   Internet:      ljdickey@math.UWaterloo.ca
                  ljdickey@math.waterloo.edu
   UUCP:          ljdickey@watmath.UUCP
