Re: [emacs-bidi] Hebrew characters categories

[TAKAHASHI Naoto]

"Yair Friedman (Jerusalem)" writes:

>> Composition is independent of categories.  You can compose characters
>> without using categories at all.  Never mind those categories designed
>> for Thai and Lao; they were defined because they help composition in
>> those languages.  You do not have to use them for Hebrew.

> Yes, but they *do* help in composition for Hebrew as well (this is why I
> chose them in the first place).  In addition, the phonetic categories
> helps in resolving placement collision.

OK, then it is reasonable to take such an approach.

> Do you think there is a better approach?  I'm sorry, but I don't know
> enough about Emacs internals to start a whole new design.  The Thai
> solution did seem like a good place to start.

I'm sorry, but I don't know enough about Hebrew to start a whole new
design.  :-)  I will read an introductory textbook again.

>>> What is the approach taken in Arabic?

>> For positioning, no new categories are defined.  Since Arabic vowel
>> signs and other marks are aligned to centre, all we have to worry is
>> whether they are placed above the base character or below.  Such
>> vertical positioning is automatically done by the display engine,
>> based on the glyph information stored in the font file.

> So do you force the vowel sign to same vertical position always? Even
> in cases such as Bah Vs Tah with Fatha and possibly Shadda? What about
> Alif?

No.  The vertical position changes.

> Do you use special presentation forms or somehow adjust the
> composition.

No.

> Do you have assumptions about the font structure?

Yes.  If

  1. the font used by our Emacs has the property
     "_MULE_RELATIVE_COMPOSE" and its value is A

and

  2. a glyph has a descent value B, where B < 0 and -B > A

then such a glyph is put _above_ the preceding character when
composed.  In case of Arabic, FATHA is such a glyph.

We also added extra white pixels above the BAH, TAH and other glyphs
so that we can get appropriate distance between the base glyphs and
FATHA.

-- 
TAKAHASHI Naoto
address@hidden
http://www.m17n.org/ntakahas/