Re: RFC: scim and gnome 2.18

Daniel Glassey wrote:
I'd like to start discussion about proposing SCIM[1] to be included in
gnome 2.18 as the default IME so that languages/scripts that need more
than xkb layouts are supported by default.

Also, scim doesn't have the xkb restriction where you can only have a
maximum 4 keyboard layouts to swtich between.

After years of being irritated at how unnecessarily complicated input methods seem to be, I finally started doing something about them for GTK+. I started a tiny version of a system we have been using internally for the last 10 years. At the moment it works fine with all applications using text widgets that allow input methods to be selected. I have tested it with some 2.6, 2.7, and 2.9 versions of GTK+.

The idea is that Linux should have something as simple and useful as Tavultesoft's Keyman available. This is a start.

1. At the moment, the code has to be compiled with the other input modules in GTK+.

2. Currently, ALT_R+SHIFT_R switches between input methods. This will be configurable later.

3. Input method files are encoded in UTF-8 and can easily be made to work with any keyboard.

4. There is no limit to the number of input methods available in each window.

5. Input methods for Arabic, Armenian, Georgian, German, Greek, Hebrew, Hindi, Lao, Russian, Serbo-Croatian (Latin & Cyrillic), Japanese T-Code, Thai, and Vietnamese are already available and working, and I have a lot more I will be converting from our old system. An example Vietnamese VIQRI input method can be seen at:

6. There is no support for sophisticated Hanzi, Kanji, Hanja, and Chu Han input, but simple table-based input for these will be easy to implement. I have some Pinyin, ZhuYin, WuBi, etc. table methods that will be added soon.

7. I'm currently working on a popup list of input methods to choose from to avoid having to cycle through all the possibilities with ALT_R+SHIFT_R.

8. A visual keyboard and visual keyboard layout creator are planned.

9. A future improvement will be to make this approach smaller for embedded systems. It doesn't use much memory now, but it can easily be slimmed down even farther. The code is currently less than 100K and I hope to keep the final release with all the tools under 250K.

MIM (My Input Methods) is *almost* ready for end users. I am adding support for a user-specific config file and code for the Han table-based input methods (which present multiple candidates at one time).

Anyone interested can email for a copy of the source as it is now.
Mark Leisher
Computing Research Lab              We find comfort among those who
New Mexico State University         agree with us, growth among those
Box 30001, MSC 3CRL                 who don't.
Las Cruces, NM  88003                 -- Frank A. Clark

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]