Newsgroups: comp.sys.apple2
Path: utzoo!utgpu!utstat!philip
From: philip@utstat.uucp (Philip McDunnough)
Subject: Re: How to Choose a Programming Language
Message-ID: <1991May12.204831.7394@utstat.uucp>
Organization: U of Toronto Statistics
References: <16117@smoke.brl.mil> <1991May11.230407.1038@utstat.uucp> <16133@smoke.brl.mil>
Date: Sun, 12 May 1991 20:48:31 GMT

In article <16133@smoke.brl.mil> gwyn@smoke.brl.mil (Doug Gwyn) writes:
>In article <1991May11.230407.1038@utstat.uucp> philip@utstat.uucp (Philip McDunnough) writes:
>>The "funny symbols" have been available for virtually every computer for a
>>long time.
>
>That's certainly not true.  None of the usual code sets (ASCII, EBCDIC, etc.)
>include the APL symbols, and even most typesetting fonts used with laser
>printers do not include them.  There have been occasional special-order
>computer terminals such as the Tektronix 4015 with an APL character set,
>but most terminals have not provided APL symbols.

While it's true that most terminals have not had support for the APL
character set, there have been several suppliers of APL terminals. As far
as laser printers using them, both STSC and Spencer(APL6800) support
printing APL code. Possibly you are using 20 years as a benchmark for a
long time.


>>The main complaint was not the unavailability( since there were
>>even ascii equivalents before Iverson came out with J) but the so-called
>>unreadability of APL code, and the difficulty people had in maintaining
>>other people's programs. This difficulty can be traced to most programmers'
>>poor grasp of the concepts of mathematics and their inability to think in
>>abstract ways.
>
>In my experience the real difficulty lay in the abysmal lack of
>professionalism on the part of APL coders, who would ignore software
>engineering issues such as the need to anticipate maintenance
>requirements.  It was a rare APL program that was adequately commented,
>let alone designed in any greater depth for maintainability.

That's true in many, but not all, cases. It's not only APL programmers
who suffer from lack of foresight.

>The same flaw can occur in improper use of almost any programming
>language.  I've seen TECO macros that make APL programs look like
>Basic English.  And of course there is the infamous Obfuscated C Code
>contest, whose winning entries are almost incomprehensible.  None of
>these reflect on the language, but rather on the programmer.

>I don't know how you could have information about "most" programmers.
>Do you really know that many people?  The best programmers I know of
>certainly think at a high level of abstraction.

I work in an 80,000 student university and have been doing it for 14 years.
I've seen a lot of programmers pass by. Have a look at the mathematics'
component of your typical undergraduate honour's degree in Computer Science
and you will see what I mean. Of course there are exceptions. There are
not that many first rate programmers. The best you refer to constitute a
some proportion of people programming for a living.

Philip McDunnough
University of Toronto
philip@utstat.utoronto.ca
[my opinions,...]
