TOPICS OF INTEREST:

Abstract:

Our natural language processing research group has taken the decision of adopting Unicode for all its data. Consequently in the process of rewriting one of our already existing applications, namely a graphical tool for tree-banking with parsing aids, we had to integrate the support of Unicode. Originally, the application in question was intended for English (1byte) and for Japanese (2-byte) only. The new specifications included a redesign as a client/server application. We "killed two birds with one stone" by using Java. This eased both the use of Unicode and the implementation of the client/server communication.

The introduction of Unicode allowed us to simplify existing C code on the server side because only 2-byte code had to be adapted to Unicode and the old one for 1-byte had just to be thrown away. The universality of the tools integrated in the server is a new feature implied by the use of Unicode: it is a valuable investment for the future, when we shall deal with Korean, another language that we intend to deal with in our research.

Another benefit of rewriting our application as a client/server application with Java is that all input methods on the client side now support Unicode. As a consequence, the interface has been alleviated of the task of managing character inputting and partly layout. Hence, the management of a new language like Korean for input/output does not necessitate a recompilation of the client code.