Natural Language-Based Software Analyses and Tools for Software Maintenance

摘要: Significant portions of software life cycle resources are devoted to program maintenance, which motivates the development automated techniques and tools support tedious, error-prone tasks. Natural language clues from programmers’ naming in literals, identifiers, comments can be leveraged improve effectiveness many tools. For example, they used increase accuracy search tools, ability navigation recommend related methods, raise other analyses by providing access natural information. This chapter focuses on how capture, model, apply conceptual knowledge expressed both linguistic information as well programming structure semantics. We call this kind analysis Language Program Analysis (NLPA) since it combines processing with extract for source program.