[Uim] Determine the semantic of reset
suzhe at tsinghua.org.cn
Fri Jun 10 08:14:50 EEST 2005
Yes. I think your description is correct.
In SCIM, the first API is focus_out (), and the latter is reset (),
does it make sense?
Kenichi Handa wrote:
>Sorry for the late response on this matter.
>Instead of quating your mails and propose my opinions
>against each of them, I'll show, what I think, an ideal
>behaviour of an application/IM-engine/IM-module.
>When a user moves an input spot or a focus out, an
>application calls the corresponding IM-engine function
>(API). The function asks the IM-module to react on that
>event. The IM-module will commit some text to the
>IM-engine. Then the IM-engine commits the text to the
>application. It's up to the IM-module how to change the
>internal status on that event. Some may reset the full
>status. Some may reset the status only partially.
>When a user or an application want to reset an input method
>to the initial state, an application calls the corresponding
>IM-engine function (API), and the function asks the
>IM-module to reset the status. The IM-module silently reset
>the status (i.e. without committing anything).
>Now I referred to two APIs. The formar is mandatory. The
>latter is, I think, optional because it should be the same
>as destroy the current input context and create a new input
>context. So, it's up to an IM-engine to provide it or not,
>but if it is costly to destroy and create an input context,
>providing the faster/lighter way is better.
>If an IM-engine provides both APIs, I believe the term
>"reset" is more suitable for the latter. Otherwise, calling
>the former as "reset" not ideal but acceptable.
>handa at m17n.org
More information about the uim