Re: project proto proposals

[Gregg Reynolds <gar at arabink dot com>]

Tarik FDIL wrote:
> Dear Gregg,
>
> My answer is about your first suggestion.
>
> I think we should focus on Linux/Unix, since :
>
> - There is no need to support arabic in other operating systems, but in linux 
> there is a lot of work to do. So, Linux should remain our primary 
> developpement environnement. Now if our products could be easily adapted to 
> other operating systems, whye not ? 
>
> - Binary portability is impossible. Full source portability is an illusion. 
> You should always adapt your program to a concrete operating system, library, 
> tools, etc. So, first lets work on linux and then adapt for other 
> environnements if needed. 
>
> - 100% of arabeye's projetcs are designed for linux/unix. So the day when we 
> would have 50% of our projects running on other operating systems, we could 
> discuss the proposal of expanding Arabeyes'scope. 
>
> UMHO of course !

Hi Tarik,

I pretty much agree with all you've said, although I'm not sure why you 
say there is no need to support Arabic in other OS's.  I guess you mean 
system-level support.  But my proposal is not to move away from 
Linux/Unix (and btw *BSD), but to expand the focus.  I think of 
development as only one part of the picture.  If I were developing 
something for Arabic, I would no doubt do so on a *nix platform, but I 
would also try to use libraries (e.g. QT) that are portable to any 
platform.  Portable libraries are available for just about anything one 
wants to do.

But Arabeyes doesn't look like only a development site to me at least. 
Lots of the work is translation, which is of course portable.  There's 
also the issue of basic support - what programs are available to do 
things in Arabic?  How do I obtain and install them?  etc.

As far as source portability is concerned, there are many many open 
source packages that have been ported to many OS's, including windows. 
I can do almost all my computing at work using open source software, 
even though I am forced to use Windows as my OS.  Open Office, the GIMP, 
Inkscape, even X by using cygwin - the list goes on and on.

My own view is that designing software for an operating system gets it 
backwards.  We should design for the community of users, not for any 
particular machine or OS.  The OS is rapidly becoming a commodity 
anyway; for me it is almost irrelevant.

Expanding the focus of Arabeyes should not inhibit work on *nix projects 
by any means.  People will always work on the things that interest them 
anyway.  What I have in mind is simply appealing to a larger audience - 
people who can't run *nix but can run open source applications on 
Windows or Mac OS X.  After all, it's a lot easier to transition to *nix 
if you know that the apps you use will be available.  And it's a lot 
easier to impress a Windows user if you can show them running code - 
e.g. the GIMP or Bayani - that solves a problem in Windows as well as 
the proprietary solution does.

I guess all that is my rather verbose way of saying that I think there's 
a big opportunity to be helpful both to more users (Windows users who 
are unaware of the availability of FLOSS stuff on windows) and to 
developers who want their stuff to support Arabic on all platforms. 
Part of that is just marketing - if such people see "The Arabic Unix 
Project" they'll probably think Arabeyes has nothing to help them, which 
I don't think is true.

(Yet another of my many personal projects for which I never have time is 
to document how to make your Windows machine look like a FLOSS platform, 
via cygwin and various other FLOSS software.  In the past few years a 
rather large number of applications that originated in the *nix world 
have been ported to Windows, and the release of QT under the GPL means 
that lots more are on the way.  Part of the idea being that you can have 
the same "system image" no matter what your OS is.  That's the main 
reason I make this proposal, actually.)

-gregg