|
The definition of user demands is one of the most important factors in the initial stage of the software development life cycle. However, traditional and currently and often adopted practices of collecting user demands such as interview, inspecting existing documents and observing current operating procedure are all just based on the viewpoints of user sides. In case of the situation that users do not have enough domain knowledge on the software system to be developed, the collected user demands will be incomplete. Therefore, if project managers or system analysts can utilize the historical projects whose functional requirements are similar to the project to be developed for the reference to design the requirements questionnaire, it can be used as an initial means of collecting user functional requirements and hence help ensure the collected requirements more precise, complete and efficient. This thesis proposes a natural language process-based software requirement questionnaire framework. It automatically derives the functional requirement items via the natural language process of the software requirements specification (SRS) of the completed software projects. Each of all requirements items are stored in the requirement questionnaire database with its associated keywords. After the establishment of requirement questionnaire database, the request for proposal (RFP) of the software system to be developed is then analyzed by the natural language process for deriving its keywords. Through these keywords, the functional requirement items can then be obtained from the requirement questionnaire database and the software functional requirement questionnaire is then developed as an initial tool of the requirements collection. This thesis also verifies the feasibility and benefit of the proposed natural language process-based software requirement questionnaire framework by the implication of an empirical case study.
|