Re: Another --> Re: GtkMovementStep of GtkTextView

[Pablo Saratxaga <pablo mandrakesoft com>]

Kaixo!

On Fri, May 25, 2001 at 10:21:37AM -0700, Chookij Vanatham wrote:

> will not want to type another key (usually enter key) to say "commit"
> (like in japanese/chinese).

That is because those language require an interactive user choice among
a list of possibilities. That is only needed for input of chinese chars
(and would be for other possible ideographic systems). A "phonetic" based
system will never require that.

> Also, we don't want another input pop-up window
> for this input method.

Again, that need for chinese and Japanese is because of the need to present
a list of possibilities.
Note that this is also due to a limitation from some toolkits (which is not
the case of Gtk, which is very good in that aspect).

> As long as the Thai character keyed/typed,
> it will be submitted to the logical steam (text), right away and the
> output method will take care it.
> 
> For my opinion, all other indic scripts would like to have the same
> behavior.

Any language/input not needed character substitution should indeed be
straighforward. the most complicated are those that need to change a char
previously entered (like telex input that needs to replace a char at 2, 3 or
4 position before the cursor position)

> The reason I say that, Thai is derived from the similar root
> of languages as indic does.

The electronic encodage and the way to type are however different (Thai
encoding is glyph oriented, indic encodings are letter oriented)

> So, if input method that is being to be used in
> indic scripts and require another keystroke to say "commit", I'm not sure if
> that's really what indic wants.

No, indic input doesn't need that; it has, however, some special keys to
have ligatures or to avoid them; but that is typed in the same text flow 
(and those special chars are stored in text file)
Do a search for 'iscii91.pdf', it is a very informative document.

> I'd like to point out that after Chinese/Japanese input method being
> commited from users by typing the commit key or somehow selecting the
> kanji/chinese characters they want from either kata/hira/pinyin/...,
> those kanji/chinese characters will be seen as ONE character in the
> text, that's why when users type <backspace>, users don't care whether
> how many kata/hira/pinyin/... characters forming those kanji/chinese
> chacters,

No, they want them seen as one character because they *are* one character.
A chinese character is not 'composed' of any kana, pinyin oranything like
that. The complex input method is need just because it is easier than
having keyboards with 1500 keys plus 8 pedals to be able to type all
posible chars (in the old days of mechanical typewritting there were such
things (well the number of keys and pedals were probably different,
I don't remember) and operators needed several years of learning.

A chinese char is not composed of what was used for input; the same that
if you type Thai in transliteration the thai chars are not composed of ascii
latin letters.


-- 
Ki ça vos våye bén,
Pablo Saratxaga

http://www.srtxg.easynet.be/		PGP Key available, key ID: 0x8F0E4975