Re: Dotless J with stroke.

On 12/11/2007 9:34 AM, Andreas St=F6tzner wrote:
> And: =93Just names=94 and =93just glyphs=94 =96 you only make it obscur=
e that way.
>
It's been a long established principle of the Unicode Standard (and=20
10646!) to freeze names of characters, once adopted. That makes them=20
different from other types of named entities; for one, many names have=20
known defects that prevent them from being reasonable description of=20
their associated characters. Therefore, if you need an accurate=20
description of a character, in the general case, you need to use an=20
_alias_, not a formal character name.

It's been an equally established principle that characters can be=20
represented by a range of glyphs, where the range depends on the=20
character. What gets depicted in the formal document tries to be both=20
typical and neutral (avoiding idiosyncratic design choices). Finally,=20
where appropriate, a serifed, specifically times-like font is used to=20
show more details and cut down on accidental confusibility between=20
representative glyphs for different characters.

The goal with the glyphs is to be an aid in identifying the character in=20
question, in a list(!), without accidentally suggesting some feature not=20
normally found in actual font designs for it.

It's a total non-goal to make the glyph a realization of the (possibly=20
faulty) description embodied in a character name. In fact, changing the=20
glyph in this instance could be construed as casting doubts on the=20
fundamental identity of the character as being the IPA character in this=20
case.

Character coding stability requires that everything be done to maintain=20
the *identity* of what is coded at a given position. The committees must=20
be constantly vigilant against "creeping  reinterpretation" of existing=20
characters.

Yes, following these principles does lead to the need for additional=20
explanation (including the use of character name aliases), but=20
realistically, there's no alternative. So pretending that they don't=20
apply is not helpful, to say the least.

I would have thought, that, as a member of the German committee, you=20
were firmly subscribed to these long-standing principles of the=20
character encoding work.

A./