View Single Post
Posts: 102 | Thanked: 187 times | Joined on Jan 2010
#276
Originally Posted by rinigus View Post
@mikecomputing, would be great to have active and contributing members in the community. I don't think we are in position where we should just sit and wait for SFOS to get better
...

No idea whether its allowed, but its a fair question. Its just insane that we have to use proprietary bits for text prediction. When (if) I get off the navigation software development, I would be happy to help anyone to get this piece of SFOS replaced with some open-sourced solution. Just would prefer to finish what I started before first. Any takers for developing text prediction for SFOS so we can drop XT9?
I think you hit a few nails here. I am willing on the same terms as you are. So for me the OKBoard is a nice piece of software and a good foundation for same or next word predictions with the swiping as icing on top of the cake. I helped promoting OKBoard by providing the Swedish language models very early on and helped others build resources for their favourite languages. OKBoard ties in nicely with the current maliit keyboard infra. As a language technologist I already back in the 90s did text prediction. Most lessons learned are that we should avoid special solutions but build something that tie in or adds to existing open solutions. In the case with xt9 we know the infra is not too restricted, so we could replace it fully with good enough API's for our free solution.
 

The Following 6 Users Say Thank You to ljo For This Useful Post: