Re: Fwd: Re: a software library for multilingualisation

[TAKAHASHI Naoto <ntakahas at m17n dot org>]

Thank you for your responses and comments.

Ayman Hourieh writes:

> I've checked the second screenshot and the demo, vowel signs look OK to me.

Please read my explanation below.

> Keep up the good work, if I need something like this, I'll consider
> your lib as one of the options.

Thank you.  We are encouraged.

Munzir Taha writes:

> Will you please point what's wrong. I believe it's acceptable though not 
> perfect.

If you Arabic speakers say it is acceptable, then it must be
acceptable.  However, my textbooks seem to have different opinion.

1.  When a consonant carries both a shadda and a KASRA, the KASRA is
    drawn between the consonant and the SHADDA (above the consonant,
    below the SHADDA).  However, we see a KASRA put below the LAM in
    the second word in the second line.

2.  Vowel signs should change its vertical position according to the
    hight of the base consonant.  The screenshot, on the other hand,
    shows all signs at fixed vertical positions.  For example, FATHA
    attached to REH or WAW should be closer to the base consonant;
    KASRA attached to final ZAIN or final LAM should be positioned
    lower to avoid overstrike.

>> I have investigated all the opentype fonts available from the Arabeyes
>> web site. Unfortunately, none of them had appropriate GPOS features.
>> So I imagine that vowels signs have little importance to native
>> speakers.

> No, they are very important to native speakers but few of them seek 
> perfectness. If it's a font issue no problem but make sure it's not your lib 
> fault.

I will check it again.

>> Anyway, we will be very glad if you make use of our library to develop
>> Arabised application programs for Unix/Linux systems.

> I am sorry I am not a developer _yet_ but is interested in becoming so soon so 
> will you please point out what the problem that this lib solves. What's its 
> scope (Do I need it if I am going to use Qt or GTK)? Who will need it? What 
> are the other alternatives and why this one?

The current version of the m17n library covers the levels of the C
library and the X library.  We are now working on multilingalising
Toolkits (Gtk is the first target).  We develop the library primarily
for application developers who want or need to write multilingual
software.  Please consult the following page for details.

  http://tsukuba.m17n.org/m17n-lib/outline.htm

Of course there are multilingual libraries other than ours; for
example, Pango for display, IIIMF for input, GNU libiconv for code
conversion.  The m17n library covers all of these aspects in a single,
unified manner.

The ICU (International Components for Unicode) libraries also covers
many aspects of multilingualisation.  While the m17n library intends
to provide better building blocks for multilingual Man-Machine
interface, ICU seems more suitable for non-interactive text
processing, like academic researches.  Which to choose is, of course,
up to you.

> Also, when I tested from the Demo on the website it seems as if it doesn't 
> conform to Unicode BIDI. If I type Arabic then English, it doesn't behave as 
> should be, at least AFAIk.

Would you be more specific?  What characters in what sequence did you
use?  The m17n library uses the fribidi library for BIDI reordering,
so we believe its result is reasonably reliable.

Best regards,

P.S.  If you reply, please CC: to mule-aist at m17n dot org so that I can
share the information with my colleagues.

-- 
TAKAHASHI Naoto
ntakahas at m17n dot org
http://www.m17n.org/ntakahas/