Re: Writing Indian Languages on GNU/Linux

[Arun Isaac]

> I am planning to start translating Guix manuals. I will first start with
> the cookbook, as its a smaller version, and can be translated faster I
> think, but If you suggest me to translate guix manual first, I will do
> so.

Translating the guix domain[1] (the command line UI messages) would be a
lot smaller and more manageable, while still being a challenging and
significant achievement. The guix manual is quite enormous. Also, the
benefits of translating the guix domain are more immediately relevant to
users and their daily activities.

> I don't have much experience in writing Devnagari or Indian languages on
> GNU/Linux. I used to work with Google Input tools, which has some phonetic
> writing. If I type "Aniket" It would be "अनिकेत in my language". So it was
> phonetic, is there anything like this on GNU/Linux? I have heard of iBUS
> but haven't used it yet. So that I won't have to learn a new keymap.

Do you use emacs? If so, emacs has its own input methods (such as
devanagari-itrans and devanagari-inscript) for Devanagari. You can
switch input methods using M-x set-input-method and M-x
toggle-input-method.

For other non-emacs applications, you'll have to use ibus or
uim[2]. When last I checked a few years ago, I couldn't get ibus to work
correctly on Guix. Things might have improved now. For non-emacs
applications, I use uim instead of ibus.

I recommend the non-phonetic input methods such as
devanagari-inscript. Phonetic input methods slow you down in the long
run. But, phonetic methods do have an easier learing curve. Then, there
are predictive input methods such as varnam[3]. This is the closest free
software equivalent to Google Input Tools.

[1]: https://translationproject.org/domain/guix.html
[2]: https://github.com/uim/uim
[3]: https://www.varnamproject.com/

signature.asc
Description: PGP signature