Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!mailrus!uwm.edu!uakari.primate.wisc.edu!aplcen!haven!umd5!suns.UMD.EDU!jph
From: jph@suns.UMD.EDU (J. Patrick Harrington)
Newsgroups: comp.lang.apl
Subject: APL11 on Sun 3/50
Message-ID: <5761@umd5.umd.edu>
Date: 11 Dec 89 06:31:06 GMT
Sender: news@umd5.umd.edu
Reply-To: jph@suns.UMD.EDU (J. Patrick Harrington)
Organization: Univ. of Maryland @ College Park, Astronomy
Lines: 33


    After asking this group about any low-cost APL that would run 
under  UNIX  on  a Sun workstation,  I received  several  helpful 
replies outlining what is available - thanks to all of  you.  The 
only *free* APL available right now seems to be APL\11, for which 
I  was directed by L.J.  Dickey to Ken Yap at Rochester.  Ken was 
kind  enough  to make the source available as  an  anonymous  ftp 
(after  I faxed him a copy of our Unix source license - it is not 
public domain, as the code is a modification of one  contained on 
the Berkeley distribution tape). Ken said that these sources last 
worked under Suns (3.4), and made no promises about 4.0.3.
    I found that they would compile under 3.5 and 4.0.3,  but the 
4.0.3  binary crashed,  and I haven't messed further with  it.  I 
have  been  running the other binary (compiled under  3.5)  under 
4.0.3. 
    It  is certainly worth having,  as there is an "apltool" (due 
to  Ken,  I gather) that provides all the characters in a  window 
environment,  although when entering the overstrikes you only see 
the last character. It is fun to play with, and I now always keep 
the  window as an icon to be used as a calculator  - a  workspace 
with all the physical and astronomical constants I frequently use 
makes it ideal for quick computations.
    APL\11  has problems,  though.  Seems prone to  self-destruct 
when  you  make  serious errors.  But the worst  problem  I  have 
encountered  is that the domino function is hoplessly broken - it 
returns wrong and even *non-numeric* characters when operating on 
even  a  simple  3 X 3 matrix.  The source code  (a8.c  ?)  looks 
fearsome to me,  so I wonder if anyone out there has been running 
this old code,  and if so, can anyone give me a clue about fixing 
this important function?  Or even user-defined inversion function 
composed of other primitives?
                                         J. Patrick Harrington
                                         jph@astro.umd.edu