bug#36717: 25.3; greek.el: deprecated vowel+oxia combinations should be replaced with vowel+tonos counterparts

[Robert Alessi <alessi@robertalessi.net>]

Hi Eli,

On Fri, Jul 19, 2019 at 09:57:38AM +0300, Eli Zaretskii wrote:
> > > But those methods that have tonos are for 'modern' greek, whereas the
> > > ones that have oxia are for classical greek, as Basil pointed out, so
> > > perhaps thereʼs no need to change anything (unless thereʼs an edict
> > > from the Unicode people that tonos must be used even when writing
> > > classical greek).
> > 
> > Good point.  I will try to look for such an edict!
> 
> We could ask on the Unicode mailing list.  There are Unicode experts
> there, and they are quite friendly.  If someone can come up with a
> comprehensive description of our situation and the issues we are
> trying to resolve, please write to unicode@unicode.org, and ask the
> questions.

As of this writing, I am doing some research on this topic.  At least,
I went through almost all the documents that are listed here:
https://unicode.org/versions/ (in reverse chronological order), and
couldn't find any statement of Greek oxia being deprecated in favor of
tonos, contrary to what is claimed here:
https://wiki.digitalclassicist.org/Greek_Unicode_duplicated_vowels

Strictly speaking, tonos is not to be mistaken for oxia: that is for
sure, as tonos was introduced as a result of a reform to denote a
tone, namely a stress on some vowels, and not a pitch, namely a rising
and falling voice on accented vowels.  Confusion began in 1986 (from
what I learned) when the Greek government decreed that tonos shall be
the acute.  In addition to this, Unicode initially served both Greek
and Coptic on the same block, starting at U+0370.  At this time, only
monotonic Greek was served.  Later, an additional block called Greek
Extended was added at U+1F00, let alone further additions being
encoded in the Supplementary Multilingual Plane...

From what I could see, many Greek fonts originally reflected the
distinction between tonos and oxia.  But nowadays, they simply mix
them up.  This is precisely why I was not able to instruct FontForge
to take into account vowels with oxia in some of the substitution
rules of Old Standard which is a very fine Greek font (see
https://ctan.org/pkg/oldstandard): vowels with oxia were simply
missing from the Greek Extended Block, and what I did not see is a
rule that instructed to absorb vowels with oxia into vowels with tonos
as the glyphs are all strictly the same.

One question remains—and I wish to express my gratitude to all of you,
Robert, Basil and Eli: since assigning vowels with tonos and vowels
with oxia to the same code points is clearly unacceptable even if the
glyphs may be identical, is there a way to input tonos and vowels with
tonos with emacs?  I use greek-ibycus4, but if other input methods
can handle these letters, I would consider any change unnecessary.

Many thanks again to all of you!

Robert