Newsgroups: comp.lang.apl
Path: watmath!watserv1!torn!utcsri!rpi!zaphod.mps.ohio-state.edu!uunet.ca!geac!itcyyz!yrloc!rbe
From: rbe@yrloc.ipsa.reuter.COM (Robert Bernecky)
Subject: Re: SIGNUM of teaching numerical methods
Message-ID: <1992Jul23.045513.20017@yrloc.ipsa.reuter.COM>
Reply-To: rbe@yrloc.ipsa.reuter.COM (Robert Bernecky)
Organization: Snake Island Research Inc, Toronto
References: <ROCKWELL.92Jul14234800@socrates.umd.edu> <1070@kepler1.rentec.com> <ROCKWELL.92Jul18201144@socrates.umd.edu> <1088@kepler1.rentec.com>
Date: Thu, 23 Jul 92 04:55:13 GMT
Lines: 51

In article <1088@kepler1.rentec.com> andrew@rentec.com (Andrew Mullhaupt) writes:
>In article <ROCKWELL.92Jul18201144@socrates.umd.edu> rockwell@socrates.umd.edu (Raul Deluth Miller-Rockwell) writes:

I have deleted the entire contents of Andrew's reply to Raul...

I think Andrew's reply is entirely correct, based on my following of
supercomputer conferences over the past years, and my experience as an
APL designer and implementor. 

In particular:

a. APLers' view that you can always reinvent the wheel, (in fact: SHOULD 
   reinvent the wheel) is dumb. It denies competence (in the sense that
   others may have a better algorithm than you, or have more numerical
   analysis sense than you), as well as denying the idea that someone
   may have taken a LOT of time to do something which you would be ill-
   advised to do again without GOOD cause (write a Phong rendering program).

b. APLers' view that everything is dense arrays, and/or everything is
   done without globals or loops. The reality of large scale computing
   says very much the opposite.

c. APLer's view that performance is for wimps. Try computing weather 
   or climate in the Best Possible Code. Now do it again in APL. 
   See you later.

APLers who deny that the above ARE problems, or that they are SERIOUS
problems, are being naive or stupid. Take your pick.

Now, having said that, I STILL think that APL has its place in 
supercomputing. It does have, or will have, its place, in the following
sense:

a. WHEN APL code can compete in loopy-type stuff, with Fortran,
   many arguments against APL will vanish.

b. When APLers admit the existence of data bases which are updated
   concurrently from a number of places, then problems such as 
   parallel computed differential equations fall out nicely.

c. When APL is JACL (Just Another Computer Language), and you can
   call APL from Fivetran and vice versa, then the problem we face today
   of Language Arrogance will vanish.

Bob

Robert Bernecky      rbe@yrloc.ipsa.reuter.com  bernecky@itrchq.itrc.on.ca 
Snake Island Research Inc  (416) 368-6944   FAX: (416) 360-4694 
18 Fifth Street, Ward's Island
Toronto, Ontario M5J 2B9 
Canada
