W3C home > Mailing lists > Public > www-international@w3.org > January to March 2006

RE: Justification in Arabic, Korean

From: Paul Nelson \(ATC\) <paulnel@winse.microsoft.com>
Date: Wed, 15 Mar 2006 23:26:49 -0800
Message-ID: <49C257E2C13F584790B2E302E021B6F90F346798@winse-msg-01.segroup.winse.corp.microsoft.com>
To: "Martin Duerst" <duerst@it.aoyama.ac.jp>, "fantasai" <fantasai.lists@inkedblade.net>
Cc: "WWW International" <www-international@w3.org>

I would think that our standards should provide some guidance for
behavior based on common practice and then leave the UA to do what it
knows best. Some Arabic fonts may create ligatures to compress text or
may use swash forms to expand text. That type of control should be left
to UAs and fonts to work out. There are some consistent behaviors that
are often seen in publication of Arabic text...

1) books or documents with longer lines will put kashida/tatweel at one
location in the word
2) newspapers or narrow column items will typically put kashida/tatweel
at all connection points to try to improve the reading experience
3) the expansion of space needs to be controlled separate from
kashida/tatweel because legal documents want no white space expansion
and books want about 20% of expansion to come from spaces for better

Our standards should provide the necessary information for UAs to do the
best possible layout with the markup they are provided.


Paul Nelson
IE Text (Beijing)

-----Original Message-----
From: Martin Duerst [mailto:duerst@it.aoyama.ac.jp] 
Sent: Thursday, March 16, 2006 2:29 PM
To: fantasai
Cc: 'WWW International'; Paul Nelson (ATC)
Subject: Re: Justification in Arabic, Korean

At 22:43 06/03/15, fantasai wrote:

 >In this case Paul and I were discussing whether, in addition to
kashida,  >which has prescribed rules for where and how much to elongate
the  >connections, whether we needed an additional justification method
that  >stretched Arabic evenly at all connection points.

Arabic is a cursive scritp, and so stretching ideally doesn't just
happen at connection points, but may also affect the shapes themselves.

Regards,   Martin. 
Received on Thursday, 16 March 2006 07:27:40 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 21 September 2016 22:37:25 UTC