Re: Quranic Proposal - Dynamic glyph substitution with OpenType

[Mohammed Yousif <mhdyousif at gmx dot net>]

On إثنين 14 يونيو 2004 12:47, Mete Kural wrote:
> > Being an OS community I suspect you want them for
> > free as well. OK, OpenType
> > (http://www.adobe.com/type/opentype/main.html) is
> > the closest thing to Open
> > Source. From MS you can obtain lots of guidance and
> > even tools to develop or
> > adjust your own
> > http://www.microsoft.com/typography/default.mspx
>
> Thanks Tom for clarifying that the problem is not with
> rendering engines since modern rendering engines
> support dynamic fonts such as OpenType fonts.
>
> Mohammed, Nadim, and Abdulhaq, in that case you should
> be able to re-design your Quran font as an OpenType
> font and implement the behaviour that Tom is
> suggesting for tanweens and other cases. 

  Mete, designing a font that can handle all Qura'nic symbols never was
  a challenge, actually the current font we have can display the Qur'an
  in a semi-perfect manner without using any OpenType hints/hacks and
  without using any ASCII characters. (This was done by adding the proposed
  characters to the font and the result is impressive, it doesn't even need
  any changes to the current standard rendering engines.)


> As he says, 
> it seems that OpenType already handles dynamically
> substituting a single glyph for a sequence of
> codepoints. This way additional characters will not be
> necessary except for a new "chairless hamza" and a new
> logical character for trigerring variant tanween
> glyphs (are there any others?)
>
  
  As I noted above, we don't want any help in implementing
  a Qura'nic font, we want a standard way of doing this not just
  suggestions. (If you want us to go this way, we will but we need
  them to be in Unicode first).
  As far as I know, the Unicode Standard doesn't state this way of
  implementing fonts at all (substituting two characters with one using
  OpenType) and hence it's a non-standard way of doing things (although
  appreciated, but we are now talking about fixing the standard not a way
  to implement this in a non-standard way).


  And what about the small alif (8 in the proposal)?
  I think it's now clear that it's like U+06E6.

  And what about 9,10,11 and 12?
  How can the font know if the needed character is the regular
  one (the tanween on top of the alef) or the one used by the
  hard-copy of the Qur'an which is certified by the most popular
  Arabic and Qur'an scholars?
  If those are not added then a font can only be either for the
  Qur'an or regular Arabic texts which is really very inconvenient
  and against the spirit of Unicode (for an international unicode font
  that should cover all the languages, for example, how can it handle
  both the 'regular tanween on top of alif' and the one used by the
  Qur'an of the characters FD3C and FD3D wasn't there?).


> You can start here:
> http://www.microsoft.com/typography/default.mspx
>
> I would like to help too.
>

  We could use your help really, but this is out of topic.
  I will, ISA once this issue is finished, start improving the font.

>
> --- Thomas Milo <t dot milo at chello dot nl> wrote:
> > > Sorry again Tom for paraphrasing for you. Please
> > > confirm whether I represented what you tried to
> >
> > say
> >
> > > properly.
> >
> > Mete,
> >
> > You paraphrase is correct with one caveat: the
> > rendering engines are not the
> > problem. The modern rendering engine is built for
> > dynamic fonts (a concept
> > pioneered by DecoType). As a result the combination
> > of Unicode and dynamic
> > fonts is a generic solution, that can be made to
> > work in most cases.
> >

  But is it a "standard way"?
  If not it either needs to be stated clearly in the Unicode Standard or
  not adopted at all because this will open the door for more vendor-
  specific methods of dealing with things.

> > If the industry does not bother to implement
> > Unicode, it is usually in the
> > Fonts Domain. For unskilled end-users, that is
> > frustrating. But I understand
> > this group has the competence and the will do
> > improve Arabic computing, so
> > you need tools to adjust or build the necessary
> > dynamic fonts.
> >
> > Being an OS community I suspect you want them for
> > free as well. OK, OpenType
> > (http://www.adobe.com/type/opentype/main.html) is
> > the closest thing to Open
> > Source. From MS you can obtain lots of guidance and
> > even tools to develop or
> > adjust your own
>
> fonts:http://www.microsoft.com/typography/default.mspx.
>

  This is all great and good (and I will try to do that with the font
  we have) but the purpose of this proposal is to fix the shortages
  in the Unicode Standard with regard to the Qur'an. If it's not
  appropriate to add more characters then your suggestions needs
  to be stated clearly in the Unicode Standard so that we can request
  from font vendors to implement them.

> > The least you should expect from the hosting
> > operating systems that you are
> > putting your cards on is that they support the
> > concept of cross-platform
> > font technology like OpenType. DecoType is also
> > developing ACE into a
> > Unicode-based cross-platform function.
> >
> > For the end-user it should be irrelevant whether his
> > Unicode text is
> > rendered by ATSUI, OT, Graphite or ACE. Therefore it
> > is of the utmost
> > importance that the encoded characters of the
> > Unicode standard are of
> > universal, plain text relevance and not proposed to
> > address an ad hoc
> > problem of a particular font on a particular
> > platform.
> >

  To be honest, I see it the other way around, I think that doing this
  in the font itself is a workaround to address the shortages of the Unicode
  Standard.
  But that's not the problem here, if your suggestions are adopted, this
  means that a LOT of work will be needed to be done for each and every
  font out there and IMHO, this is unproductive redundancy that should really
  be avoided. (not mentioning that this is a lot of work that involves
  confusion for the user typing the text, type this and this then a special
  mark to get the needed character is really frustating)

-- 
Mohammed Yousif
Egypt