From: Philippe Verdy (verdy_p@wanadoo.fr)
Date: Sun Sep 30 2007 - 07:32:55 CST
Serge Rosmorduc [mailto:rosmord@iut.univ-paris8.fr] wrote:
> If we take the current standard for encoding hieroglyphs, the Manuel de
> Codage (which is not perfect), this would mean we would require unicode
> aware softwares to recognise things like embedded groups, with
> corresponding scaling algorithms, etc...
No. Not they would remain optional if the relations (before, above...) were
encoded as format controls: Unicode can give them ignorable properties,
their rendering is not absolutely required, but at least they can be used
for semantic distinction purpose.
If a renderer is unable to produce the intended layout, they could still
substitute them with a reasonable glyph (like the MdC ASCII punctuations).
Even the modifiers (mirroring, rotations, ...), it would be possible to
encode them as combining characters (that may have an optional diacritic
glyph rendered after the hieroglyph if they are not supported by the
renderer, despite these will cause much less problem for existing text
renderers as they imply no complex scaling or positioning, but only basic
substitution (the problem is much simpler here than with Indic scripts used
today that require complex splitting, reordering and contextual substitution
and positioning rules, notably for Tibetan where this is a very complex
problem to implement properly).
This archive was generated by hypermail 2.1.5 : Sun Sep 30 2007 - 07:37:36 CST