Re: inverse Bidi

["Chahine M. Hamila" <mch at chaham dot com>]

Nadim Shaikli wrote:

> --- "Chahine M. Hamila" <mch at chaham dot com> wrote:
> > Salam,
> >
> > I finally located the implementation of inverse Bidi in the ICU library
> > (yeah, it wasn't that hard to find, but someone had to move a few
> > fingers and strike some keys...). Thanks Behdad for pointing out to the
> > existence of a reverse bidi algorithm (even though it is *not* part of
> > the standard Unicode specification as you implied).
>
> Could you please explain what this library does in "plain" english (I
> haven't gone to their site yet to read about it) and what advantages
> does it offer (I gathered from reading your mail that its an X-coordinate
> converter of some sort) ?  It is stictly a display thing ?

basically, in very few not-so-technical words (a more detailed explanation
will come when I can dedicate some time to it), inverse bidi does what
unicode's bidi spec does with Englihs but doesn't with Arabic, i.e. add bidi
capabilities to an Arabic monodirectional text. The current approach being
either use English monodirectional and no Arabic or use (whatever kind of)
bidi, which without going into details, is not as practical for edition as
monodirectional edition, and in its unicode version is not compatible with
Arabic monodirectional texts.
"Inverse bidi" transforms an Arabic monodirectional text into a "unicode
bidi" text. This is not perfect, but is a great step forward and is more or
less a green light for going further in implementing monodirectional Arabic
support, which is much easier to handle from both the developer's and user's
point of view, and which is likely to open the gate for thousands of Arabic
software in a row through minor hacks.

If this explanation is confusing, an example without technical explanations
might help:
Take vi, write an English text. Use that same text in Word for example,
which does bidi (I don't know if it's stricly unicode compliant or not, but
the idea is there). The English text will be perfectly read.
Now, say you use Akka, which allows you to type Arabic in any existing
software in the same way and comfort you type English. Take vi, write an
Arabic text. Use that same text in Word and it won't display right, because
Unicode's assumptions on logical storage do not match it. If Word allowed
inverse bidi, that text would be displayed right.

Salam,
Chahine