Path: utzoo!attcan!uunet!cs.utexas.edu!news-server.csri.toronto.edu!qucdn!bakerj
From: BAKERJ@QUCDN.QueensU.CA
Newsgroups: comp.lang.apl
Subject: First Impressions of J
Message-ID: <90289.112513BAKERJ@QUCDN.QueensU.CA>
Date: 16 Oct 90 15:25:13 GMT
References: <9014@latcs1.oz.au>
Organization: Queen's University at Kingston
Lines: 47


In the last week I have been experimenting with J version 2.3
from Iverson Software.  So far I am impressed.  J is remarkably
complete (the documentation indicates what remains to be done).
implementation of "pure" rationalized APL.

A few brief observations.

     1)  It's amazing how fast you get used to the use of ASCII.
         J's ASCII symbols are well chosen.  The language has
         has a nice look to it and it's easiar to type and
         print than convential APL. I appreciate
         the APL character set but languages like J and Nial
         have convinced me that it is not necessary.

     2)  As a language J rigorously implements the extended
         Iverson notation first discussed in Rationalized
         APL back in the early 80's.  J has taken a number
         of much needed steps.  Standard line number based
         function definition has been thankfully trashed.
         Syntactic constructs that don't fit into the
         rational scheme (ie index brackets) have been
         eliminated. Verbs (functions) in old Sharp APL that had
         inappropriate function ranks have been fixed.
         Many new verbs, adverbs and conjunctions have
         been added.  J is certainly powerful.  It's well beyond old
         pre-nested array APL's.  It's in the same class as
         APL2 and Nial.

     4)  The ISI interpreter written by Roger Hui has not
         implemented workspaces.  He has adapted an ASCII
         script scheme that is certainly sufficient for
         for demonsration software. I've been impressed
         with the performance of the interpreter and it's
         scope.  This is the only $24.00 package I now of
         that computes complex matrix inverses.

     5)  Overall ISI's J is an amazing bargin.  For $24.00
         you get the J interpreter.  Iverson's paper (ISI J Dictionary),
         and a little booklet called -- Tangible Math.
         I have a feeling that J is going to attract
         a lot of interest within and,  to a lesser degree,
         outside of the APL world.   Who knows it may even
         revitialize this inbreed little corner of computerdom.


                                    John Baker