Re: Symbols used in a computer language

Nicholas Bodley wrote:

>APL is a computer language that is distinguished by quite a few unique
>symbols. (See the consecutive code points starting at U+2336; the Unicode
>3.0 Names index lists other code points.)
>One ref.: <http://www.users.cloud9.net/~bradmcc/APL.html>
>
>Although "language" in computer programming has attributes different from
>those of a human, spoken (or gestural) language, nevertheless, it does
>share some attributes.
>
>I only wanted to call the attention of the Qalam community to a
>non-trivial-sized set of unique characters used only in a form of
>communication very different from that of ordinary human languages.
>
>APL seem to be uniquely concise; the internal processes that take place in
>response to executing only one symbol seem to correspond to a
>quite-significant amount of code in other languages, in some instances.
>At one time, there was a portable computer that had an APL keyboard, and
>executed the language more directly than usual.
>
>Does some computer code written by talented programmers have some of the
>attributes of poetry to those (few?) who know both? I suspect that such
>could be said. An elegant algorithm, expressed in a lower-level language,
>at least, can have a beauty, I think; that sort of beauty is akin to what
>mathematics (mostly "higher"?) can offer in some instances, it seems to me.
>
>

I've programmed in APL, and indeed, it is beautifully concise. I recall
sometimes typing in a line of code and then just pausing to admire its
elegant beauty (really!) It's happened in other programming languages
now and then, but APL is something else. My standard example of APL
involves imagining a program that takes an arbitrary number of numbers
for input and prints out the average. A small matter of a few lines in
almost any language. Eight *characters* in APL. I did an assignment
for an AI class in APL (my classmates struggled with huge programs in
LISP) that was maybe seven lines long--only because the first six were
comments. It's been said that if you can do it in APL, you can do it in
one line of APL.

A friend of mine advocated writing what he called "obscene APL", in
which the overriding rule was avoidance of parentheses if at all
possible (and sometimes even if not possible). This was particularly
fun, since APL has no algebraic rules of precedence: unless overridden
by parens, all expressions are evaluated strictly from right to left
(yes, I said right to left). The classic example was a little
expression to compute sqrt(A^2 + B^2).

After I've blathered on and on with that topic... what does this have to
do with writing systems? It's true that APL has some weirdo characters
(I even remember one that isn't coded in Unicode and probably shouldn't
be: to break out of a program when it's asking you for input, you type
"O<backspace>U<backspace>T". Many of the other symbols were originally
assembled with backspaces too (quote-quad was quad<backspace>quote,
etc), but what else need be said about it? Poetic programming is a
matter for another list.

>[Getting 'way off topic, but: Write 11 33 55, "doubles" of the first odd
>integers. Rearrange: 113 355. Put the 355, the bigger one, on top.
>Divide, 355/113. The result is a remarkably-good approximation to [pi]. No
>better approximation with fewer than 11 digits, total, exists.
>(That is 104348/33215.) To me, this is beautiful, and not higher math. at
>all.]
>
>

There's a cool pi approximation with a gematria of a qere/ketiv pair in
the Bible somewhere, but I think it gets the approximation just slightly
worse than 355/113, and you wish it could be worked to get the 355/113
version... OK, that is *almost* writing related.

~mark