Abstract: User input in the form of image data is received from a user via a natural user interface. A vector difference between an adjustment start position and a current position of the user input is calculated. The vector difference includes a vector position and a vector length. The vector position is compared to stored rail data, and the vector length is compared to a stored threshold length. The rail data describes a plurality of virtual rails associated with an application. Based on the comparisons, the user input is matched to one of the plurality of virtual rails and a notification describing the matching is provided to the application. The application, thereupon, transitions from a first command to a second command corresponding to the matching virtual rail without receiving any explicit termination gesture for the first command from the user.

Abstract: User input in the form of image data is received from a user via a natural user interface. A vector difference between an adjustment start position and a current position of the user input is calculated. The vector difference includes a vector position and a vector length. The vector position is compared to stored rail data, and the vector length is compared to a stored threshold length. The rail data describes a plurality of virtual rails associated with an application. Based on the comparisons, the user input is matched to one of the plurality of virtual rails and a notification describing the matching is provided to the application. The application, thereupon, transitions from a first command to a second command corresponding to the matching virtual rail without receiving any explicit termination gesture for the first command from the user.