From: Kenneth Whistler (kenw@sybase.com)
Date: Tue Jul 14 2009 - 13:52:08 CDT
Asmus,
> > Typographically, it might sit just slightly to deep. However, being a
> > spacing character, I wonder whether this is not just a glyph/font issue.
> For some characters the "it's just slightly off" really means it's a
> different character.
It this case it does not, however.
You and everybody else are arguing character encoding based on
a well-known book that is a typographical one-off -- and a notoriously
bad example of typography, at that. This book was typeset by a
printer (U.C. Berkeley Press, 1968) which couldn't even insert
typeset Chinese characters -- they were gapped, and then written in by hand.
> The character we are trying to analyze is clearly a subscript. In the
^^^^^^^^^
The glyph(s), rather. The identity of it as a character is
precisely in question.
> samples it harmonizes with the subscripted double prime (or double
> vertical bar?) for the tertiary stress.
Not always, because the examples vary in their typography.
But the examples in that case help make Szabolcs' case, because
the use of U+02C8 MODIFIER LETTER VERTICAL LINE and
U+02CC MODIFIER LETTER LOW VERTICAL LINE for primary stress
and secondary stress, respectively, is a well-known IPA convention
that Y.R. Chao would have been familiar with. If you look
at *that* set of characters on that page, then the low ring
should indeed be construed as the U+02F3 MODIFIER LETTER LOW RING.
(not IPA, but IPA-inspired).
The other examples are a typographer's hack for the same thing,
using a subscript letter "o" from the font.
Incidentally, the use of the period in Y.R. Chao's system
as a indication of neutral tone on a following syllable
is also arguably related to the IPA system of supresegmentals,
because in IPA, U+002E is used to indicate syllable breaks.
>
> In the samples
> (http://www.stud.uni-karlsruhe.de/~uyhc/files/images/p1070002.preview.jpg)
> it does not fully harmonize with the period used for neutral stress -
> they appear to not have the same center, as one would perhaps expect.
> However, the period is ordinarily positioned so that it aligns at the
> top of these subscripts, which gives some consistency in appearance.
This is way overanalyzed. This kind of argumentation from
one-off (bad) typography is not a good precedent for
getting more characters added to the standard of dubious
semantics and appearance -- they only add to the confusion about
which is the "correct" character to use in representing texts
like this.
> Not having actual samples of material using the low ring, I can only go
> by its appearance in the charts, and it's quite a bit lower there, lower
> than a subscripted character.
>
> As long as that chart glyph is not an aberration, I would very much
> hesitate to forcibly unify these.
And I think you would be wrong. Szabolcs is correct.
For representing this particular convention in the Y.R. Chao's
text (which isn't even a regular part of the Gwoyeu Romatzyh
romanization which sees any wide use -- because it isn't
a direct indication of pronunciation, but rather an editorial
shorthand for saying that an alternative pronunciation with
the tone neutralized also occurs), the best choice is:
U+02F3 MODIFIER LETTER LOW RING
For people who want to argue text appearance and alternative
origin theories you also have available:
U+006F LATIN SMALL LETTER O (styled subscript)
U+2092 LATIN SUBSCRIPT SMALL LETTER O
U+2080 SUBSCRIPT ZERO
U+FF61 HALFWIDTH IDEOGRAPHIC FULL STOP
U+3002 IDEOGRAPHIC FULL STOP
So argue away and do what you will -- but the *last* thing the
Unicode Standard needs is *another* low ring character encoded
based on this evidence and this usage, to further muddy the waters.
--Ken
This archive was generated by hypermail 2.1.5 : Tue Jul 14 2009 - 13:55:08 CDT
