From: Kenneth Whistler (kenw@sybase.com)
Date: Mon Mar 07 2005 - 19:50:57 CST
Dean asked:
> For my purposes I envision a set of nine characters that specify what
> areas should be subtracted from, or de-emphasized in, the glyph when rendered:
> The rules for their use are rather simple:
These rules are *not* simple, from the point of view of a renderer.
>
> a) These rendering instruction characters would precede the visible
> character they modify.
Conceptually, these are akin to formatting characters, but treated
as combining marks. They would have to follow, rather than precede
a base, to fit into rendering models for Unicode.
> b) Any number of them can be used, disallowing duplicates.
How would duplicates be disallowed?
> c} Any order is allowed, with ordering being insignificant.
This creates equivalence and comparison problems.
>
> So I have two questions:
>
> 1) What do you think from an encoding point of view?
I think this is clearly not plain text, and should not be encoded
this way. As others have suggested, this can and should be handled
via text markup.
>
> 2) Would any of the major text rendering systems even consider
> implementing such a system if it were encoded?
I seriously doubt it. Not for plain text.
> By implement, I mean, for
> example, in plain text draw all of the letter "A" as normal, except its
> lower right corner in a lighter shade to indicate damage there.
That isn't plain text. And to do this kind of thing at all, you
would need special font support. The very concept of a 3x3 grid
(or a 2x2 grid, for that matter) doesn't map well, in the general
case, to rendered glyphs for all writing systems. It already
assumes a level of abstraction that identifies some square box for
each glyph to be portrayed in, so that quadrant damage or lacunae
can be identified. How would I expect a renderer to apply that
abstraction to Arabic or Mongolian script, for example?
What *could* be appropriate for encoding as characters, from the
fields of paleography and epigraphy here, would be entire symbols
indicating quadrant damage -- in other words, some thematic take on
sets of quadrant symbols such as U+2596..U+259F, U+25E7, U+25E8,
U+25F0..U+25F3, etc, which might reflect use in text to *discuss*
glyph damage and lacunae, etc. This would be quite different from
trying to encode a bunch of format controls to actually make
the text *render* with damage and lacunae.
--Ken
This archive was generated by hypermail 2.1.5 : Mon Mar 07 2005 - 19:52:18 CST