Dear Qalamites,

While not strictly on-topic, this message (a reply to a message I'd sent) seemed
to be of uncommon interest. Mr. Hosny is one of the people who are localizing the
XO (One Laptop Per Child -- OLPC computer) software and (perhaps) firmware.
Perhaps it's being excessively cautious, but I have deleted his e-mail address
as a courtesy.

I fairly recently joined the Localization list at OLPC, and in response to a
question, offered a rather-long message describing what's involved in
working with RtoL, Arabic, and other RtoL scripts. I was out of date, and
the replies were most courteous; perhaps the message was not useless.

----- Original Message -----
From: Khaled Hosny
To: Nicholas Bodley nbodley@...>
Sent: Wed Jul 30 18:12
Subject: Fwd: Re: [Localization] Arabic Projects

On Wed, Jul 30, 2008 at 01:09:35PM -0400, Nicholas Bodley wrote:

> On Wed Jul 30 11:44 , Khaled Hosny sent:

[KH]
> >Arabic typewriters used what so called "Simplified Arabic script" where
> >each letter has only 2 forms (on used as isolated and final, and the
> >other as initial and medial) reducing the number of needed glyph while
> >remaining acceptably readable.

[nb]
> Most interesting! Did the Shift key select the other form for a given letter?
> For such people as students of writing systems and typography, it would be good
> to have some images of typewritten Arabic on the Web.

Yes, shift used to select the isolated/final form, unfortunately I
couldn't find such samples on the web (I should remove the dust from my
father's typewriter and make some samples).

[KH]
> >If by "mechanically" you mean metal type,

[nb]
> I did.

[KH]
> >then Arabic printing houses as early as 1800's produced very carefully typeset
> >Arabic books that we yet to have a computer system that can imitate.

[nb]
> That's just fascinating. I would not have thought it possible. I do hope that
> somebody has saved a few fonts of type from that period. One would think that
> there was quite a large number of different sorts (glyphs) in a given typeface.
> As well, one wonders whether the body of each piece was only a rectangle, or
> whether some more sophisticated scheme was worked out.

They used extensive sets of hundreds of glyphs with tens of contextual variants
of each glyph.

Here are two scans from a Mushaf printed in 1924 using metal type, this
is one of the finest metal typesetted books.
http://www.tafsir.org/vb/attachment.php?attachmentid=850&stc=1&d=1180677875
http://www.tafsir.org/vb/attachment.php?attachmentid=851&stc=1&d=1180678163
(Note that Sura heads are calligraphed while the body of the page is
composed).

[I looked up some definitions: A Mushaf is apparently the Qur'an,
bound in one volume.
[Tafsir refers to commentaries on the Qur'an. Editing the URL to substitute
a few other 3-digit numbers in place of "850" and "851" brought up some
pages, completely in Arabic, as well as .doc-formatted files
that seemed to render quite well in a recent version of OpenOffice.org for
Windows XP SP2.
[Sura, also Surah -- pl. Suar or Suwar, are chapters of the Qur'an. -- nb]]]

[nb]
> I think these details would be of interest to subscribers to the Qalam mailing
> list on yahoogroups.com. That list (only sporadically active) is for people
> interested in writing systems, and sometimes branches out a bit into linguistics
> and typography.
>
> Perhaps I could forward your message to that list, with your permission (also
> with OLPC's permission).

Of course you can :)

Regards,
Khaled

--
Khaled Hosny
Arabic localizer and member of Arabeyes.org team

(If you see duplicate[s] of this message, I apologize.
This Web mail software is not professional, although
the rest of Speakeasy's services are very good. --nb)

--
Nicholas Bodley
Waltham, Mass.
Sent from Speakeasy.net web mail



[Non-text portions of this message have been removed]