[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Greetings from new member
- To: doc at arabeyes dot org
- Subject: Greetings from new member
- From: gar <gar at arabink dot com>
- Date: Sat, 20 Nov 2004 10:32:33 -0600
آهلا وسهلا
After lurking for quite a while (first discovered arabeyes a few years
ago, if I remember correctly) I've decided to take the plunge and try
to find the time to get involved, so this is my introductory missive.
First off, congratulations to everybody involved; arabeyes is a very
impressive project and it has progressed quite a bit since I first
discovered it. Obviously computing is still in its infancy in the
Arab world, and it's exciting to see a community coalescing around a
project like arabeyes with its emphasis on free software and full
arabization.
As for myself, I'm more-or-less fluent in Arabic (enough so to land a
job working with Arabic after years as a software developer) and I
hope to help out mainly by assisting the translation effort.
I've got a million ideas for arabizing software (who doesn't?). I've
been thinking about it for many years, ever since I first started
studying Arabic and realized the potential of the computer in making
this arduous task easier. But also (and more important) the need to
close the technology gap between the "advance" world and "the rest" of
the world is a great motivator.
In particular, here's what I see as a fundamental issue: the
nurturing of an indigenous, monolingual software development
community. By this I mean it should be possible for a kid anywhere in
the Arab world to learn to hack code without having to learn English
first. A simple measure of success is easy to define: successful
implementation of a major software project entirely in Arabic. (The
same obviously goes for any linguistic community; I just happen to
have learned Arabic so that's where I concentrate.)
So I propose as a possible goal for arabeyes the design and
implementation of an Arabic language programming language. Virtually
all programming languages today are optimized for expressiveness in
English; what would happen if we had a language optimized for
expressiveness in Arabic? Arabic is terrifically powerful and
flexible when it comes to abstraction; who knows what interesting
things might come out of a project to create an Arabic-based
programming language?
As a start, I propose the arabization of Scheme, probably using the
Guile implementation. Scheme is very simple, very powerful,
relatively easy to implement, and it's a wonderful language for
learning computational concepts. I think it would be relatively
simple for a good experienced programmer to modify a Scheme
implementation to work with Arabic keywords and handle Arabic strings.
Wouldn't it be fun to write code in Arabic?
Another goal I propose, for which I have an immediate pressing need,
is the arabization of a basic set of XML tools. First off, a decent
editor with a user-friendly interface. I've downloaded and tested 20+
editors in the past few weeks, and have yet to find one that supports
Arabic. Many of them, strangely enough, almost support Arabic - they
shape the glyphs properly, display Arabic right-to-left, but editing
doesn't quite work; the cursor behaves strangely, so you can't tell
where the true insertion point is. In any case, I suspect it wouldn't
take all that much effort to enable basic Arabic support. Possible
open source candidates are http://www.butterflyxml.org and
http://www.xerlin.org. There are other free xml editors but their
names escape me at the moment. Once I've finished my evaluation of
editors I'll post the results to arabeyes.org.
In addition to XML editing, I need a good file comparison tool that
handles Arabic. I've looked at a bunch of them but still haven't
found one. If there isn't a good one out there I suppose the thing to
do would be to arabize gnu diff.
Finally, just to stoke a little discussion, I'd like to get this off
my chest: Unicode is a catastrophe for the Arabic-speaking world and
for any community using a right-to-left writing system. Why?
Because contrary to the claims of the Unicode designers (and virtually
everybody else, alas) Arabic is not bidirectional, nor is any other
language in the world. The alleged evidence for said bidirectionality
is numbers in Arabic. But in fact this claim is based on spoken
Arabic, which is obviously not the same as written Arabic, and
furthermore is not uniform - it is not uncommon to hear Arabic
speakers utter numbers in LSD-first order. This is incredibly
frustrating, because the issue is so simple - all you have to do is
realize that written Arabic is a monodirectional language with LSD
polarity - it writes numbers Least Significant Digit first. The fact
that people usually utter numbers MSD first is totally irrelevant to
the writing system. English happens to be a monodirectional language
with MSD polarity. Big deal. I don't see anybody claiming that,
because we say "seventeen" and not "tenseven" the number should be
written "71". It's idiotic, the bidi requirement. grrrr.
Think of it this way: Unicode is a hidden tax on the entire rtl
world, whose value I would guess must surely reach into the billions
of dollars. Without the bidi requirement it would be a relatively
simple matter to adapt the billions and billions of dollars worth of
software out there to rtl languages. With the bidi requirement, most
of it will never be adapted. Just try asking your favorite small
software company to support Arabic and you'll see what I mean: "it's
too expensive to implement bidi".
And that's not even considering the bizarre (and IMHO totally
unacceptable) behaviour of the cursor in bidi enabled software. Nor
the many other issues that make Unicode Arabic one of the worst
designs in the history of technology.
Whew! Got that out of the way. Thanks for listening. ;)
Gregg Reynolds