원문정보
초록
영어
The currently available approaches for processing natural language (NL) software requirements specifications are semi-automatic and require user intervention. Moreover, these approaches result in less accurate and imprecise object oriented software models. Recent research in the area attributes the informal nature of natural languages for less accurate analysis of software requirements. On the basis of this axiom, we have identified that direct translation of a natural language to a formal language is the actual problem. In this paper, we propose that instead of direct translation of a natural language to a formal language, first, we will transform the natural language text to a semi-formal language that is not only simple and easy to translate to a formal language but also provides with higher accuracy. We have incorporated Structured Business Vocabulary and Rules (SBVR) language as the semi-formal medium in natural language to object-oriented models. The presented approach automatically generates the object-oriented software models from natural language software specifications using SBVR as a pivot representation.
목차
1. Introduction
2. English to SBVR Transformation
2.1 Processing English Software Specification
2.2 SBVR Vocabulary Extraction
2.3 Generating SBVR Fact Types
2.4 Applying Semantic Formulation:
2.5 Applying Structured English Notation
3. Object-Oriented Analysis of SBVR Rule
3.1 Extracting Classes:
3.2 Extracting Instances:
3.3 Extracting Class Attributes
3.4 Extracting Class Methods
3.5 Extracting Associations
3.6 Extracting Generalization:
3.7 Extracting Aggregations
4. Conclusions
References