Tuesday, 6 August 2013

When is a programming language "more like English"?

When is a programming language "more like English"?

I an tasked with introducing a non-programmer who is a very good
web/graphics designer to Rebol.
I am faced with my own claim that Rebol and Smalltalk are "more like
English" than Ruvy and Python for simplw scripting.
My claim looks good when compare simple web scraping : rebol read and
write trump Ruby read and write for URL's and files.
But then I come to the function confirm.
Confirm can also be confirm/with
But it seems obvious to an outside that confirm would be "more like
English" if it accepted at least three equivalent optional refinements :
confirm/options confirm/choices confirm/with
Some might even thing /from and /using belong in there to be "like English".
The Curl web content language, which I use more often than rebol (
although I use some rebol to generate that Curl markup) has some inherited
quirks with optional constructors that take a string. Once those oddities
are in the language, they are hard to iron out.
But change need not "break legacy code" if the languages of "the gentle
curve' accepted equivalencies just as it does, say, new being optional
with instance creation constructors.
Suppose a language such as Curl were to catch on big in western China,
central Asia, Malaysia, Indonesia, Africa - wouldn't the interest o those
many new users far out-weigh the preferences of the few established users
? Wouldn't it be more rational and altruistic to offer the many cosistency
WITH flexibility rather than to resist change?
Some such transformations can be seen in Python across major releases and
may prevail in Ruby. But in those widely used languages (relative, say, to
Rebol and Curl) many voices were heard over the years rather than a few
"cranky/grumpy" voices.
Indeed, when I made a suggestion at one "new language" forum, I was
immediately barred by the forum mediator as an obvious "troll".
But as I already work in two expression-based languages and Smalltalk
dialects/implementations and Prolog dialects (including, Horrors, prologs
with Types and type declarations), how could I be a "troll" for raising an
issue in a language forum where the language itself was still evolving and
not yet adopted by even a single major corporate or gov IT Dept. (unlike
Curl and even Icon) ?
Just learn it and get used to it and if it is STILL a problem to you, then
maybe ...
English orthography changed not just because of dictionary authors and
editors but also because of preverse Classics men ... pedants who have
bequeathed spellings that only make some little sense in a nation with
Latin as the second language and Greek as the third.
An analogy could be drawn with a few recent progamming languages and their
peculiarities which at least one now seeks to shed. But will it be "more
English-like" and why should that impress a Japanese or an Italian or a
... ???
I am probably too old now to stop seeing Ruby as just a poor way to do
Smalltalk in a world where RAM and DASD are no longer what limit us. Who
are the C programmers today who need Objective-C to transition to GUI
programming ? What is Perl to a pythonic visitor to Ruby-land ? And no,
not one word about Clojure.
If the keyboard evolves, even the argument that [] are less trouble than
{} may vanish ... so why not make the delimiter choice optional ? Is one
"more English-like" ? Neither are like the wide variety available in
Japanese ....

No comments:

Post a Comment