Community
Participate
Working Groups
Currently it uses ICU4J BreakIterator but this has to change.
I don't think this will be possible that easily. If we want proper word boundary detection for all possible languages, I don't think there is a way around using a dedicated Unicode library such as ICU for this.