gnome-applets-charpick 2.22.1-alt1

Now the groups of characters are made up according to similarity of glyphs/sameness of the base character.

But such grouping is not convenient at all for real use. A user usually works with a certain language (or languages). So, it would be convenient if the characters were grouped according to the languages for which they are used. 

For example, there should be one group for the umlauts used in German. Then, if a user happens to type German words one day, he would select that group (that's just one action) and work more quickly with the applet. (Now, he would need to constantly switch between the groups.)

On implementation: fontconfig already has a database of languages' orthographies, so gnome-applets-charpick could borrow this database.