More Latin-9 input methods?

More Latin-9 input methods?

[Kai Großjohann]

Would it be a good idea to create more Latin-9 input methods?  For
example, Germans are probably switching from Latin-1 to Latin-9, but
the german-prefix input method produces Latin-1 characters.

Maybe an input method german9-prefix could be created.  (And it
should have support for entering the Euro sign...)

Opinions?

(Once Emacs switches to Unicode, this might become a moot point.
With unify-8859-on-decoding-mode switched on, it's probably a moot
point already.  Hm.)

kai
-- 
Silence is foo!

Re: More Latin-9 input methods?

[Eli Zaretskii]

> From: Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai =?iso-8859-1?q?Gro=DFjohann?=)
> Date: Fri, 03 May 2002 22:01:52 +0200
> 
> Would it be a good idea to create more Latin-9 input methods?

Yes, I think so.  Ideally, I think there should be a set of Latin-9
input methods that exactly mirrors Latin-1 input methods, so that
Latin-1 users migrating to Latin-9 won't feel any significant
differences in their default configuration.

Re: More Latin-9 input methods?

[Richard Stallman]

    Would it be a good idea to create more Latin-9 input methods?  For
    example, Germans are probably switching from Latin-1 to Latin-9, but
    the german-prefix input method produces Latin-1 characters.

I think we should improve input method support for Latin-9,
but I think that it would be much simpler to add a flag
to make Latin-1 input methods handle Latin-9 instead,
rather than making a whole set of alternate input methods.

Perhaps unify-8859-on-decoding-mode is that flag.

Re: More Latin-9 input methods?

[Kai Großjohann]

Richard Stallman <rms@gnu.org> writes:

>     Would it be a good idea to create more Latin-9 input methods?  For
>     example, Germans are probably switching from Latin-1 to Latin-9, but
>     the german-prefix input method produces Latin-1 characters.
>
> I think we should improve input method support for Latin-9,
> but I think that it would be much simpler to add a flag
> to make Latin-1 input methods handle Latin-9 instead,
> rather than making a whole set of alternate input methods.

There are characters in Latin-1 that are missing from Latin-9, and
vice versa.  If there is only one input method, what should it do
with these characters?

(As a very concrete example, the Latin-9 version of german-prefix
should allow for entering of the Euro sign, which is not in Latin-1.)

kai
-- 
Silence is foo!

Re: More Latin-9 input methods?

[Eli Zaretskii]

On Sun, 5 May 2002, Kai =?iso-8859-1?q?Gro=DFjohann?= wrote:

> There are characters in Latin-1 that are missing from Latin-9, and
> vice versa.  If there is only one input method, what should it do
> with these characters?

Insert them nonetheless (in the only possible charset)?

Re: More Latin-9 input methods?

[Richard Stallman]

    There are characters in Latin-1 that are missing from Latin-9, and
    vice versa.  If there is only one input method, what should it do
    with these characters?

Perhaps it could have a list of pairs of characters, each Latin-1-only
character paired with one Latin-9-only character.  Then there would be
a sequence for each pair.

When you type this sequence and the flag is set to Latin-1, you
would get the Latin-1 character.  When the flag is set to Latin-9,
you would get the Latin-9 character.

Re: More Latin-9 input methods?

[Kai Großjohann]

Richard Stallman <rms@gnu.org> writes:

>     There are characters in Latin-1 that are missing from Latin-9, and
>     vice versa.  If there is only one input method, what should it do
>     with these characters?
>
> Perhaps it could have a list of pairs of characters, each Latin-1-only
> character paired with one Latin-9-only character.  Then there would be
> a sequence for each pair.

It might be difficult to pair them up.  For example, the code point
for the latin-1 character that looks like "3/4" is the same as the
code point for the latin-9 character that looks like Y with two dots
on top.  So one is a letter and the other is a nonletter.

Is it really a problem to have the duplication between Latin-1 and
Latin-9?  In the long run (after Emacs changes to Unicode internally)
it would be good to make the distinction go away.  But in the
meantime, Latin-1 and Latin-2 also share a lot of characters, yet
the input methods are separate.

kai
-- 
Silence is foo!

Re: More Latin-9 input methods?

[Richard Stallman]

Latin-1 and Latin-2 are very different.  Each has a number
of single-language input methods, but they are for different languages.
By contrast, Latin-1 and Latin-9 are meant for the same languages,
are they not?

Meanwhile, most of the input methods (the ones for specific languages,
suchas german-prefix) use only characters that are the same in Latin-1
and Latin-9.  So it would be trivial to set these up to work in both
character sets depending on a flag.  Meanwhile, we should
have separate input methods latin-1-postfix and latin-9-postfix.

Re: More Latin-9 input methods?

[Kai Großjohann]

Richard Stallman <rms@gnu.org> writes:

> Latin-1 and Latin-2 are very different.  Each has a number
> of single-language input methods, but they are for different languages.
> By contrast, Latin-1 and Latin-9 are meant for the same languages,
> are they not?

I confess that I'm not sure.  Latin-9 adds some letters which look as
if they might be intended for an Eastern European language, but I'm
not sure.

> Meanwhile, most of the input methods (the ones for specific languages,
> suchas german-prefix) use only characters that are the same in Latin-1
> and Latin-9.  So it would be trivial to set these up to work in both
> character sets depending on a flag.  

I guess that each Latin-9 method would be equal to the Latin-1 method
plus a way to type the Euro sign.  So they are indeed very similar.

Let me try to cons up something.

> Meanwhile, we should have separate input methods latin-1-postfix and
> latin-9-postfix.

These are already there, which is good.

kai
-- 
Silence is foo!

Re: More Latin-9 input methods?

[Eli Zaretskii]

> From: Kai.Grossjohann@CS.Uni-Dortmund.DE
> Date: Tue, 07 May 2002 16:06:08 +0200
> 
> > By contrast, Latin-1 and Latin-9 are meant for the same languages,
> > are they not?
> 
> I confess that I'm not sure.  Latin-9 adds some letters which look as
> if they might be intended for an Eastern European language, but I'm
> not sure.

Those characters were added not for Eastern European languages, but
for Estonian, Finnish and French.  (Amazingly enough, 3 French
characters are not in Latin-1.)

Other than that, Latin-1 and Latin-9 indeed cover the same languages,
with the exception of Cornish (which Latin-1 covers while Latin-9
doesn't).