{"items": [{"author": "Danner", "source_link": "https://www.facebook.com/jefftk/posts/365638380125796?comment_id=365673176788983", "anchor": "fb-365673176788983", "service": "fb", "text": "I agree that swype could be a ton better at understanding with this type of thing. dasher's method of considering probability of all possible next letters/words would be a decent step to a general solution to your goal, just by feeding enough text into dasher, it learns the probabilities and could give you a decent shot.", "timestamp": "1331650960"}, {"author": "George", "source_link": "https://www.facebook.com/jefftk/posts/365638380125796?comment_id=365717520117882", "anchor": "fb-365717520117882", "service": "fb", "text": "Are you familiar with Dasher?", "timestamp": "1331656795"}, {"author": "Jeff Kaufman", "source_link": "https://www.facebook.com/jefftk/posts/365638380125796?comment_id=365777016778599", "anchor": "fb-365777016778599", "service": "fb", "text": "@George: I am, but I didn't know it existed for phones.  I just tried it and it's definitely slower than Swype even though it may have a better language model.", "timestamp": "1331664302"}, {"author": "George", "source_link": "https://www.facebook.com/jefftk/posts/365638380125796?comment_id=365778846778416", "anchor": "fb-365778846778416", "service": "fb", "text": "Slower as in runtime? Or slower at completing words? I think Dasher uses a PPM language model.", "timestamp": "1331664574"}, {"author": "Jeff Kaufman", "source_link": "https://www.facebook.com/jefftk/posts/365638380125796?comment_id=365781530111481", "anchor": "fb-365781530111481", "service": "fb", "text": "@George: slower as in how long it takes me to enter text on my phone with my fingers.", "timestamp": "1331664909"}]}