Can the new beaming algorithm handle issue #11 gracefully? Currently, the result is still incorrect. Werner