Re: Latin-1's apostrophe, grave accent, acute accent

From: Kenneth Whistler (kenw@sybase.com)
Date: Mon Aug 02 1999 - 17:24:31 EDT


Constantine Stathopoulos said:
>
> On 2/8/1999, at 11:34 рм, schererm@us.ibm.com wrote:
>
> >i believe that there is not much of a question: not doing the 'right' thing
> >because of improper implementations of a 7-bit standard is _bad_ for all the
> >reasons that you gave.
> >go for B, the straight apostrophe.
>
> Which, i.e. the 'straight' apostrophe glyph, is quite wrong as far as the
> Greek script is concerned. In Greek - where the symbol and its name comes
> from - apostrophe is unambiguously depicted as a homeoglyph (or homoglyph
> in many fonts) of PSILI (U+1FBF).
>
> In fact, I am very curious to see how this one will be resolved as far as
> international Unicode fonts are concerned.
>

People will no doubt argue mightily about this for the next decade,
and then just as abruptly forget what all the fracas was about, and go
on to the next thing to argue.

Actually, for Greek, the solution is well in hand, I think.

ISO 8859-7 implementations for Greek already have the homeoglyph for psili
available at 0xA2. That is mapped to Unicode U+2019 RIGHT SINGLE QUOTATION MARK,
which is also the character that the Unicode Standard recommends for use
as an apostrophe (meaning specifically the right comma-shaped form, as opposed
to the straight up form of U+0027). Just as Latin-based implementations will
have to gradually migrate from use of U+0027 (or CP1252 0x27) for use of
apostrophe in such words as "don't" to use of U+2019 (or CP1252 0x92), so
will Greek-based implementations migrate from use of U+0027 (or 8859-7 0x27
or CP1253 0x27) to U+2019 (or 8859-7 0xA2 or CP1253 0x92). Word-processing
applications are making or have made this transition. Other applications
will gradually catch up -- particularly if we keep making consistent
decisions about how to steer out of this morass as more system and application
support for Unicode becomes available.

Option B is the right way to go.

--Ken



This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:50 EDT