Automatic test-input generation is an important and expensive testing activity that greatly impacts the effectiveness of automatic testing. There are unique challenges to automatically generating test inputs for Android applications (apps) due to the unique mechanism of the Android platform. Therefore, numerous automatic test-input generation methods for Android apps have been proposed. This study proposes a description framework to demonstrate the key issues in automatic test-input generation and includes three dimensions to describe the technique (representation of search space and the generation and evaluation of candidate test inputs) and two performance metrics for the dimensions (thoroughness and efficiency). Furthermore, existing techniques, as well as potential future work, are discussed.