Fwd - RE: Why Arabic Shaping?

[Nadim Shaikli <shaikli at yahoo dot com>]

I'm cleaning my mailbox and thought I'd archive this message on
arabeyes' General list (for completeness).

No reply/comment needed.

 - Nadim

--- Marco Cimarosti <marco.cimarosti at essetre dot it> wrote:
> Subject: RE: Why Arabic Shaping?
> Date: Tue, 11 Sep 2001 13:57:54 +0200
> 
> Hi, Nadim.
> 
> I read over again the old thread on the Unicode list, and my impression is
> that you concluded it with no questions asked.  So, I don't see which reply
> you are waiting from me.
> 
> Moreover I don't understand what you mean when you say "stand-alone"
> ISO-8859-6, and this seems a key concept in all your message.
> 
> In that thread, I have often played the "devil advocate", i.e. I tried to
> see the situation from points of view that do not correspond to my opinions.
> 
> So, I will repeat some of the key points of my point view, and I suspect
> that you won't agree with one or more of these points:
> 
> 1) The process of displaying on screen (or printing on paper) the glyphs
> corresponding to abstract Arabic characters should be completely enclosed
> in a module called "rendering engine".
> 
> 2) The rendering engine should be the only module which knows about glyphs
> and fonts, and no other part of the system should be allowed to access fonts
> in any way.  Better said, fonts are *part* of the rendering engine.
> 
> 3) ISO-8859-6 does *not* encode character shapes: it only encodes abstract
> letters.  How to select from a font the exact glyph(s) to drawn text on the
> screen is out of the scope of ISO-8859.
> 
> 4) Similarly, Unicode does *not* encode character shapes: it only encodes
> abstract letters.  How to select from a font the exact glyph(s) to drawn
> text on the screen is out of the scope of Unicode.
> 
> 5) Additionaly, Unicode contains some characters representing glyphs for
> Arabic letters and groups of letters but, generally speaking, these
> characters should *not* be used.  They are included for easing round-trip
> conversions of old character sets.  Occasionally, they can be used
> internally as helpers for a *VERY* *NAIVE* implementation of a rendering
> engine.
> 
> 6) It is *NOT* *POSSIBLE* to standardize the mechanism by which abstract
> Arabic characters are converted to glyphs.  There are different graphic
> traditions and techniques, which are based on different principles and yield
> completely different rendering.
> 
> 7) There is *NO* *NEED* to standardize the internal functioning of rendering
> engines, including format of fonts and the mechanism by which glyphs are
> stored and extracted in fonts.  This does not mean that such a
> standardization is not *possible*.  Such a thing is certainly possible, in
> theory, and perhaps it would also be useful but, again, it is *NOT*
> *NEEDED*.  Different systems may use totally different techniques and
> display Arabic in totally different ways, as far as their input is the same
> kind of "abstract characters" (ISO-77596 or Unicode), and their output is
> considered "readable" by users.
> 
> Ciao.
> Marco
> 
> > -----Original Message-----
> > From: Nadim Shaikli
> > Sent: Tuesday, September 11, 2001 04.05
> > To: marco.cimarosti at essetre dot it
> > Subject: Re: Why Arabic Shaping? [posted on Tue Aug 28, 2001]
> > 
> > 
> > Marco, should I expect a reply from you on the "Why Arabic 
> > Shaping?" thread, or should I proceed.  I really don't see
> > the need for a stand-alone ISO-8859-6 (its not being shipped
> > anywhere - linux/solaris/etc) due to its short-comings and
> > uselessness (note "stand-alone").  I'll proceed with 10646
> > (include 8859-6 and Form-B)...
> > 
> > Alas, if I've exhausted you and the list - please let me 
> > know.  A "no comment" reply would suffice.
> > 
> > Thanks and sorry to pick on ya personally...
> > 
> >  - Nadim


__________________________________________________
Do You Yahoo!?
Yahoo! GeoCities - quick and easy web site hosting, just $8.95/month.
http://geocities.yahoo.com/ps/info1