Control the load of input methods

On 2011-05-25 01:00, Kyle Nitzsche wrote (on bug #663776):
> Also, I think this whole *side* conversation highlights a *critical*
> need for clearly articulated expectations for Ubuntu input methods on
> a multiuser system with multiple languages. Someone should own the
> job of developing and articulating those expectations.

I'm afraid that I don't see the possible problem here, Kyle. Could you
please elaborate?

No changes were made in Natty in order to adapt the im-switch/ibus behavior to the model for language related settings that was introduced in Natty. The reason is that such changes were considered too intrusive for an SRU. In Oneiric some code changes for that purpuse has been made in language-selector. There proved to be no need to alter the im-switch or ibus package.

This is how it works on an updated Oneiric system:

* Through im-switch it's determined which input method system that is automatically loaded at the session start. The choice is made out from the user's set display language.

* As long as a user has never made any changes, the default for the CJK languages is 'ibus', while the default for most other languages is 'none'.

* On the "Language" tab in the Language Support UI the user can set a value that overrides the default for the currently selected language, i.e. the first language on the list.

* For instance: If somebody, who speaks simplified Chinese, prefers and has selected English as the display language, but still wants to be able to input things in Chinese, s/he just changes the im setting from Language Support. The new value is stored as a symlink in ~/.xinput.d, and will be the im choice for English at session startup as long as it's not changed again.

* You can start the ibus daemon manually whenever you like, if a need unexpectedly comes up in the middle of a session.

Personally I think this makes sense. Please let me know of any concerns you may have. I change the status of this bug report to "Incomplete", which means that it may expire in a couple of months if no more comments are submitted.

Thanks! Personally this looks perfect to me.

Thanks for saying that. There were a few misconceptions on the other bug, probably due to the mix of two topics on the same bug report.

Hopefully others are fine with this as well. Time will tell.

Closing this bug, since there proved to be no need to discuss the topic further.