From: Kenneth Whistler (kenw@sybase.com)
Date: Wed May 05 2004 - 15:36:19 CDT
Dele,
> This mail is written with the Yoruba Keyboard that was rolled out yetserday.
And the good thing to note here is that the characters you
entered have survived in your email client, been sent
on the internet, entered and exited a Unix-based mail server
hosted on the Unicode web server, were redistributed to
hundreds of other people, survived *their* email clients,
and are correctly interpreted there.
> Please just look at the issue raised earlier raised.
>
> ÁÉẸ́ÍÓÚ
> áéíẹ́ọ́ú
>
> Looking at the above it is obvious that the acute on top of the e and o with
> dot below is a bit too high almost to the point of looking like a cedilla
> under E.
As John Cowan noted, that is not at all obvious. How they look will
depend on what font you are using and what renderer. When I look
at them in IE using the Andale Unicode MS font, the acutes on the
e-dot and o-dot are just perfect, but the acute accent on the E-dot is,
unfortunately, placed too low.
Such anomalies can be fixed with ligature tables in fonts designed
to work for languages such as Yoruba, Edo, etc. They don't require
changes to the *character* encoding.
>
> In transit the acute and the grave could be removed by just putting the
> cursor in between ẹ́ and ọ́ because ther are combined in a way that is not
> binding.
In transit the "e" between the "D" and "l" in your name, Dele, could
be removed by just putting the cursor in between the "e" and the "l"
and backspacing.
The character *encoding* is not the technical means whereby one
guarantees that those things that should stay together stay together.
It is implemented software on top of the character encoding which
does so.
See Sections 5.11 Editing and Selection, 5.12 Strategies for
Handling Nonspacing Marks, and 5.13 Rendering Nonspacing Marks,
in the Unicode Standard for what the standard has to suggest regarding
how software should be implemented to deal with such questions.
>
> It even becomes a compounded problem during copying and pasting because the
> accent occupy two cursor space.
Whether or not a combining character sequence is made visible to
an end user in terms of keyboard input or cursor behavior in
editing is a result of how the software is implemented.
> I still think with all these observations
> something must be done.
Of course, but what must be done is to specify how the software
should be implemented to accord with the intuitions and expectations
of Yoruba and Edo speakers and writers, and then to work with
people to ensure that the software *does* what is expected.
It is erroneous to expect that adding precomposed characters to
the Unicode Standard will accomplish that. You (or somebody else)
could spend the 2 to 3 years required, in an attempt to get these
characters added as precomposed characters, and even if they
*were* added, in the end, 3 years from now, you'd be faced with
exactly the same implementation problems in practice, as
normalization rules would force those precomposed characters into
the exact same combining character sequences you have already
entered into this email. You'd be no further along is resolving
the *real* problems, which are now in the fonts, keyboards,
and word processing systems.
Regards,
--Ken
>
> Dele Ọlawọle
This archive was generated by hypermail 2.1.5 : Fri May 07 2004 - 18:45:26 CDT