About ibus 1.3.7 ibus languages

Aron Xu happyaron.xu at gmail.com
Wed Sep 22 17:27:12 UTC 2010

Hi Shawn and all,

On Mon, Sep 20, 2010 at 07:57:12PM -0000, Computerguy wrote:
> I was wondering if you wanted to use ibus-m17 for all of the other
> languages in ibus 1.3.7. I have contacted martin pitt and told him for
> japanese at least using anthy is better than ibus-m17. Anthy for
> japanese has a lot more features than ibus-m17.
> Thank you, Shawn
> -- 
> This message was sent from Launchpad by
> Computerguy (https://launchpad.net/~blah-679)
> using the "Contact this user" link on your profile page
> (https://launchpad.net/~happyaron).
> For more information see
> https://help.launchpad.net/YourAccount/ContactingPeople

I think there are some misunderstanding for the ibus packages and what
we are doing. Please let me try to make it clear.

Here are some facts:

Firstly, ibus is almost used only by CJK users AFAIK, that is to say
Chinese (simplified and traditional), Japanese and Korean. I believe the
number of users in each of the four are descending sorted in this order
as well. 

Secondly, ibus-m17n was in CD before this time we see it is being
removed. This package provides some input method that are not good
through m17n libraries. 

Thirdly, the better choices for users:
 * Chinese: ibus-pinyin (ibus-chewing is better for some traditional 
            Chinese users, but ibus-pinyin is still good enough comparing 
	    with ibus-m17n)
 * Japanese: ibus-anthy
 * Korean: ibus-hangul

Question: What we are doing?

Now we are removing ibus-m17n (and dependencies) from the CD to save
some space, because it is not good. Maybe people will complain that with
ibus-m17n there is at least a working one but without it there is none.
Yes, but removing it should be still considered a step forward rather
than regression, because we can find better way for aid of input method
in near future's development. 

Question: Shall we also consider it a regression since some user cannot
          input characters when they just finished a fresh installation?

Definitely not. People who don't use an input method may think there is
a one that can use is better than none. But things seems to be contrary
in this specific field, because the low quality of those input methods
will do harm to user experience and when there is none users will know
they need to install it (install language support or simply a package).
In fact, we may consider it a big step forward, as explained below.

Question: Why removing ibus-m17n a step forward? 

We make Ubuntu for best user experience, so we are urging to provide top 
class applications we can. As I've said before, ibus-m17n provides not 
good input methods for users. According to our user's feedback (Chinese 
users, via forum, mailing list, irc, etc), newcomers are often be 
confused by those pre-installed input methods and thinking they are the 
recommended ones (because Ubuntu will make a best choice for their product 
if the vendor of Ubuntu is sane). Unfortunately, the input method there 
are really hard to use: for example, two "pinyin" in ibus-m17n. Input 
methods on Windows (R) platform are evolving quickly, the user will soon 
be frustrated by such 1990s style of input method in ibus-m17n, then drop
Ubuntu because they feel the experience is UNBELIEVABLE BAD. Removing
ibus-m17n from default installation is a best way of solving the

Question: What for users who need ibus-m17n?

I've wrote at the very beginning of this letter, that most people who
are using ibus are CJK users, and these users have better choices that
we can install for them when they are installing language support. For
those who really wants input method provided by ibus-m17n, they can
simply install it handily.

Aron Xu
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 490 bytes
Desc: Digital signature
URL: <https://lists.ubuntu.com/archives/ubuntu-devel-discuss/attachments/20100923/8de8d442/attachment.pgp>

More information about the Ubuntu-devel-discuss mailing list