Chinese input methods: Overview and comparisons

The widespread use of modern-day electronic devices with Chinese text input functionalities has allowed Chinese to be typewritten and exchanged at an unprecedented rate. To serve these growing needs, continuous efforts have been made to improve the performance of Chinese input methods, which come in three main types: root-based, stroke-based and sound-based. In this paper, the precursors of the present day Chinese input methods, including some standard indexing systems of sinograms (i.e., Chinese characters) and transliteration schemes of spoken Chinese, are first introduced, followed by some of the most popular methods that exemplify the three types. Then, general comparisons between these three types of methods are provided. In particular, the encoding efficiencies of nine methods in current use are evaluated quantitatively using a set of four criteria, and it is found that root-based methods perform the best of the three types of methods in this regard. The practical values of the other two types of methods should not be diminished, however, especially because stroke-based methods are easier to learn than root-based methods, while sound-based methods can relieve the user from the need to retrieve the visual forms of sinograms. The paper concludes with considerations over the implementation aspects of Chinese input methods, from which are generalized several pointers for future developments.