Newsgroups: comp.lang.apl
Path: watmath!watserv2.uwaterloo.ca!watserv1!seary@sfu.ca
From: seary@sfu.ca
Subject: Notation on the Net
Message-ID: <9304102000.AA02783@monashee.sfu.ca>
Originator: daemon@watserv1.uwaterloo.ca
Sender: news@watserv1.uwaterloo.ca
Organization: University of Waterloo
Date: Sat, 10 Apr 1993 20:00:52 GMT
X-Mailer: ELM [version 2.3 PL11]

I'm new to Unix and internet, and I think comp.lang.apl is great. I
read it to keep up with product announcements, for the theoretical
discussions, to see how people solve/code various problems, and not
least because as an APL programmer, I mostly work alone (APL allows
non-trivial programming to be done this way):
 "There are no words to express the abyss between isolation and having
one ally...two is not twice one; two is two thousand times one..."
(actually, it's not that bad; I know quite a few casual APL users).
I haven't thought I had much to add until the recent discussion about
the lack of APL vs J input. So here goes...
I think there are two questions here:
  Is J notation an acceptable substitute for APL notation on the net?
  Is J defining the direction that APL should go in future developement?
I only want to tackle the first one, thanks.
The argument seems to be that since internet only supports (7? 8?)-bit
ASCII, we need a substitute for conventional APL notation in order to
express algorithms, trade code, etc... So we have to use J.
Two problems with this:
  J has developed away from even Dictionary APL so that it is no longer
an acceptable substitute. It can still be enjoyable to translate some algo-  
rithm expressed in J (Vx.y) into APL[2 or II], but in some cases it
gets a bit mind-bending. Even J users have problems with this one:
witness the amount of discussion on how to use } (amend: now different enough
from Dic APL that I gave up trying to translate Hui's great "..uses of
{ and }..." paper.) Wait til they get to |: (cant: now also
different). I suspect this is behind some of the controversy: if one
could just translate J into APL (automatically?), who could complain?
To anticipate responses: please don't get me wrong! I like J. I have
all the publications; Hui's "Implementation..." should be used as a
course on C; J *IS* pointing at least *SOME* directions; it is also
completely fluid. And I know that APL[2 and II] has similar problems, but
they are getting smaller...
   Second, (begging the question dept.) do we need a substitute? I
took a look at a couple of other newsgroups where one would expect
notation to be a problem: sci.math.* (a completely unscientific sample from
among thousands of articles). What I saw was a pastiche of:
   TeX (e.g., \int for integral)    2     2
   typographical tricks (dy/dx, or x  +  y , or u    - u       ,etc...
                                                 [n]    [n-1]
   fragments of C, FORTRAN, Pascal (undoubtedly containing typos)
   LOTS of pointers to articles in print, articles in TeX or PostScript
    camera-ready format, or executable code (the last two with internet
    addresses for ftp-ing).
   NO discussion of the need for a universal ASCII-based notation. 
    (I will suggest a reason for this below). If you like getting lots
    of mail, you might try suggesting this to one of these forums... 
  
  So there you have it...some of what we have been doing all along
 (except for the last bit). As for the last bit, must we insist that
 the code fragments we send over internet be *EXECUTABLE*? It can usually
 be made clear (by context, which is why I hesitate to give examples) what
 one means using a pastiche of typographical tricks (oh, all right:
  +/"c[2]A     NB. sum each rowsof A -- I know it's a stupid example)
  or by using TeX-style escape sequences (I've seen this used here, too).
  If there is a substantial piece of code, point to its ftp address --
  and avoid getting flamed for misusing bandwidth. How to store for
  ftp-ing? ZIP the .ATF (now supported by Manugistics as well). It's
  pretty much the WSIS of yore. (YORE ERROR if you use the original).

    Speaking of misusing bandwidth...
    Comments?
                                 A.

