This study deals with the problems and the ambiguities
encountered in Thai software requirements specifications, the
nature of the Thai language and a simple approach. We
evaluated our approach starting from the questionnaire. The
results show that our approach can generate syntactically
correct sentences. The resulting sentences are understood and
accepted by the clients and the software engineers with high
accuracy. We have developed this approach for developing a
tool for the future. Our approach tries to help the software
engineer to write syntactically correct requirements and also
avoid ambiguities that can lead to misunderstanding between
the clients and software engineers.
In future work, we plan to extend our approach so that
more complex rules will be handled. We will also look into
developing a semantic network using conceptual modeling.
We believe our study is a useful guideline for writing software
requirements specifications in other natural languages too.