From: verdy_p (verdy_p@wanadoo.fr)
Date: Wed Nov 05 2008 - 22:19:37 CST
> Message du 05/11/08 17:59
> De : "=?utf-8?Q?António MARTINS-Tuválkin?=" <antonio@tuvalkin.web.pt>
> A : unicode@unicode.org
> Copie à :
> Objet : Re: apostrophe plus acute accent (side by side, as a single diacritic placed above C and G)
>
>
> On 2008.11.04, 20:42, Benjamin M Scarborough
> <benjamin.scarborough@student.utdallas.edu> wrote:
>
> > I would think that <c, U+0313, U+0301> or <g, U+0313, U+0301> would
> > give you the diacritic you want, but I haven't heard of any
> > side-by-side diacritic positioning rules described for Latin, so it may
> > render as an acute _above_ a comma, and I'm sure that isn't what you're
> > looking for.
>
> What about U+0063 U+0301 U+031B? Character U+031B COMBINING HORN has
> combining class 216 [Above right attached].
using two diacritics that have distinct combining classes diables their stacking. In your example the combining
horn is attached (so it cannot stack) and the U+301 accent would still stack normally above. In addition they are
specified on distinct horizontal positions (accent centered by default, harn on the right).
Anyway he was describing apostrophe (which is normally written near above right when it is not a diacritic, but the
standard says that the normal apostrophe characters MAY be used as diacritics.
So I would simply write <x, u+0301, U+2019> (if using a curly right apostrophe, which normally looks like an
elevated comma). However this apostrophe will not be displayed above and treated as a combining diacritic like
U+0313 (comma above).
Now what he wants is to write the apostrophe before the acute accent, in a way similar to what Greek does with its
soft spiritus: you'll note that even Greek does not always place the spiritus and accents above the letter, but
near their left, when the base letter is capital. I'm not sure about what Benjamin is trying to do but it seems
that he wants to transliterate Greek to Latin while keeping the Greek accents and spirits. If so, he should not
call this an "apostrophe": what he is seeking is a soft of Latin version of these Greek diacritics (however Greek
has no such diacritics on its gamma letter and has NO equivalent for Latin letter C, so I am wondering which
language Benjamin is trying to write).
Unicode and ISO 10646 have unified the diacritics used between Latin and Greek (and also Cyrillic). If you look at
what Greek fonts are doing when there are two "above" diacritics on the same base letter, they are in fact NOT
stacking them, even if they have the same numeric combining class.
»» The numeric class is just a HINT about the position, but exceptions are made depending on base letters and other
surrounding diacritics, so that the vertical stacking CAN be replaced by horizontal alignment if needed.
Nothing can prohibit a Latin font to do the same as with Greek letters, provided that the relative order is
logically kept: the first encoded diacritic MUST be on the left of the second encoded diacritic (instead of above)
if those combining diacritics have the same NUMERIC combining class, and the pair may also be displayed above left
or above right of the letter instead of above it for readability (like with Greek capitals), and it may even be
placed so that one diacritic will be placed on each side (this highly depends on the font style and notably the
traditional line-height/M-height ratio, and tradition like in Greek).
Look for other examples with the Vietnamese Latin letters that can have two diacritics normally combining above:
notably a circumflex and a acute or grave accent: they are typically NOT stacked but displayed side by side (but
vertical stacking is still permitted and is used in "Times New Roman"), unlike with "Verdana" that uses side by
side presentation.
But if you look more closely at the Verdana font, you'll see that the diacritics are noteffectively side by side
but the accent is still slightly above the circumflex, but kerned and offseted a bit in the free space left by the
circumflex, just because it can fit there. I think we are in such a situation here: it's not an encoding problem
but a question of font design, and specific adaptation of pairs of diacritics with some base letters in specific
fonts.
This archive was generated by hypermail 2.1.5 : Wed Nov 05 2008 - 22:22:53 CST