On Wed, Nov 30, 2011 at 09:47:31PM +0100, Keith J. Schultz wrote:
> Hi Khaled,
> 
> I am afraid you are completely misconceived about what the subject matter is 
> here.
>       1) adding a diacritic mark(glyph) is composing a a glyph, You are able 
> to output it
>           on its own.
> 
>       2) There is a difference between the glyph ä and adding the diactirc 
> mark of umlaut to
>              an a. 
>              a) in the first case you can use two different colors because it 
> is one glyph. 
>              b) in the second you can use two different colors because you 
> two glyph in order
>                 two compose the ä glyph.
> 
>       So we are talking about composed glyphs whether you realize it or not.

No we are not:
1) adding a mark to a glyphs is one thing, and composing them into a
   single glyph is another thing (in Latin fonts the later is a common
   result of the former, but that is mere implementation detail). The
   fully vowelled Arabic word مُحَمَّدٌ have a mark (or even two) in each
   character but no font will compose each <base><mark> into a single
   glyph (not a font build by a sane person). The same is even true for
   versatile Latin fonts that allow proper positioning of arbitrary
   <base><mark> combinations (e.g. many fonts from SIL).
2) the OP is developing his own font, which is neither covering Latin
   glyphs nor using precomposed glyphs.

Regards,
 Khaled

> 
> regards
>       Keith.
> 
> 
> Am 30.11.2011 um 13:56 schrieb Khaled Hosny:
> 
> > On Wed, Nov 30, 2011 at 11:10:11AM +0100, Keith J. Schultz wrote:
> >> Hi All, 
> >> 
> >> I jump back in. I will cite anybody because what has been said is correct.
> >> 
> >> But,
> >> 
> >>    1) trying to compare a browser, XeTex engine and LuaTeX will not help
> >>        as they have different methods of composing their output.
> >>        That is how they compose and position their glyphs.
> > 
> > As far as OpenType processing is concerned they all should give the same
> > result, else there is a bug somewhere.
> > 
> >>    2) Most important a composed Unicode glyph is supposed to be just one 
> >> color!!
> > 
> > No one talked about composed glyphs (certainly not the OP), it was just
> > a marginal and unrelated issue to the problem being discussed; coloring
> > combining marks without breaking OpenType mark positioning.
> > 
> >>    3) Once you start using color a Unicode composed glyph you no longer 
> >> are positioning
> >>         a single composed glyph, but two or more glyphs.
> > 
> > So? Again, coloring components of composed glyphs is not what is being
> > discussed here.
> > 
> >>    3.a) color designed in TeX at.al is designed be applied to a box and 
> >> not glyphs!! 
> > 
> > I've hard time understanding what this mean or how the difference, if
> > any, is material. AFAIK, TeX knows nothing about color, it handled just
> > like any other driver special which TeX makes no effort to interpret not
> > to mention "applying" it.
> > 
> >> The question remains how to position the composed glyphs and where and how 
> >> the color attribute
> >> is added to the output. 
> >> 
> >> There are therefore two solutions:
> >>    1) The Tex way:
> >>            create a macro to compose the glyph and do the positioning and 
> >> coloring.
> >> 
> >>    2) The developer way:
> >>            change the engine so that it firsts generates the composed 
> >> glyph and then goes back
> >>            and then applies color to the different glyphs. 
> > 
> > 3) The post-2000s way; use/build an OpenType font with proper combining
> > mark positioning and apply the colors to individual glyphs à la what
> > FireFox/LuaTeX and may be many other does.
> > 
> > Regards,
> > Khaled
> > 
> > 
> > --------------------------------------------------
> > Subscriptions, Archive, and List information, etc.:
> >  http://tug.org/mailman/listinfo/xetex
> 
> 
> 
> 
> --------------------------------------------------
> Subscriptions, Archive, and List information, etc.:
>   http://tug.org/mailman/listinfo/xetex


--------------------------------------------------
Subscriptions, Archive, and List information, etc.:
  http://tug.org/mailman/listinfo/xetex

Reply via email to