Re: Characters and glyphs - was Re: [I18n]bdftruncate again ;)

[Keith Packard <keithp at keithp dot com>]

Around 10 o'clock on Jan 24, Brian Stell wrote:

> A simpler example is that 'A' and 'a' are the same character, but 
> different glyphs. If Unicode was being even-handed, it would
> include one but not both of these glyphs, and application software
> would have to calculate which glyph to use each time the character
> was drawn.

One of the desirable attributes in the design of Unicode was for lossless
conversions to and from existing document encodings.  Existing Arabic
encodings didn't expose the separate presentation forms while existing
Latin encodings did.  A similar arguement can be made for han unification;
each national encoding can be converted to Unicode and back without losing
information.

Not that Unicode has no Latin biases, but I believe this particular one 
demonstrably improves it's suitability as an interchange format for 
existing documents.

Keith Packard        XFree86 Core Team        Compaq Cambridge Research Lab