Allow specifying prediction database for a language
In 1.0.5.16 Jolla silently added the possibility for the keyboard developer to choose the prediction engine for the current language. However, adding new predictions to the user databases is useful, but sometimes can pollute the other languages with fake predictions.
In my opinion, there should be another option for choosing an user prediction database for each specific language (even the non-supported ones). This will allow separating the input and probably easy backup.
Crowd-learning prediction databases might be a good feature too, though much harder to implement.
I found this very useful on my android tablet, when it detected I was not typing in English it automatically switched the language to the one I was typing in and predictions stayed fine for the language. On Jolla I am having the issue as Bundyo has described that other language words are getting mixed with English prediction database. I would like to see this feature implemeneted
Mariusmssj ( 2014-04-15 23:34:41 +0200 )editThis only happens if the languages are similar, but annoying nonetheless. :)
Bundyo ( 2014-04-15 23:35:55 +0200 )edit