From: Kenneth Whistler (kenw@sybase.com)
Date: Thu Jul 24 2003 - 21:59:13 EDT
Paul,
This position seems unduly perverse to me, and reflects an
unduly constrained notion of what "not effect the rendering"
means.
First of all, you are clear that CGJ is formally a
combining mark, and not a format control, right?
Second, many of the format controls *do* affect the rendering,
but are intentionally removed from the "glyphing stream",
since their effects may be handled by rendering logic with
is producing the glyphing stream to be painted. So it is
clear that there is not a one-to-one matching between the
class of entities which "affect the rendering" and the class
of entities with are passed to a glyphing stream for painting.
Third, space characters are funny edge cases -- in particular
ZWSP, which in most instances does not affect the rendering
(although it can). Presumably a rendering engine *could*
paint every run between spaces (including ZWSP) independently,
and never pass any space character to be painted, but I
assume that is not the way most rendering is done, and that
the font has some say in the metrics of spaces handed to it
for display. If so, why be more permissive for spaces than
for CGJ, when other considerations argue for passing it
to the font, as in the current implementation on Uniscribe?
Fourth, even though CGJ itself has no displayable glyph,
and even though it does not serve as a format control for
neighboring characters the way ZWJ and ZWNJ do, it is
clear from John Hudson's discussion that it *does* affect
rendering in an indirect way. Failing to paint the CGJ results
in a problem of attachment points, and it also makes it
more difficult for a font maker to define the "ligatures"
in the font needed to get the expected results. Making
appropriate text display more difficult (or irresoluble)
for font designers *is* affecting the rendering.
So I would urge you to think twice, and then maybe again
before unilaterally deciding to remove, based on a
philosophical principle, behavior that makes possible
a straightforward resolution of an otherwise difficult
problem for Biblical Hebrew.
--Ken
> At 11:06 PM 7/23/2003, Paul Nelson \(TYPOGRAPHY\) wrote:
>
> >It is my understanding that the CGJ should not effect the rendering and
> >is therefore should be removed from the glyphing stream. In the future
> >the CGJ will not be visible in the rendering process and therefore
> >should not be counted on to use that way.
>
> So in future the CGJ will not be painted?
>
> John Hudson
This archive was generated by hypermail 2.1.5 : Thu Jul 24 2003 - 22:30:24 EDT