Though Unified Modeling Language (UML) has been widely used in software development, the major problems confronted lie in comprehension and testing. Dependence analysis is an important approach to analyze, understand,...Though Unified Modeling Language (UML) has been widely used in software development, the major problems confronted lie in comprehension and testing. Dependence analysis is an important approach to analyze, understand, test and maintain programs. A new kind of dependence analysis method for UML class diagrams is developed. A set of dependence relations is definedcorresponding to the relations among classes. Thus, the dependence graph of UML class diagram can be constructed from these dependence relations. Based on this model, both slicing and measurement coupling are further given as its two applications.展开更多
With the scale of programs becoming increasingly bigger, and the complexity degree higher, how to select program fragments for slicing has become an important research topic. A new type of criterion called interesting...With the scale of programs becoming increasingly bigger, and the complexity degree higher, how to select program fragments for slicing has become an important research topic. A new type of criterion called interesting index is proposed for selecting parts of procedures or procedure fragments to do program slicing. This new criterion considers not only the subjective aspects in users, namely users' emphasis on the time efficiency, storage capacity or readability,but also the objective aspect in large procedures. It also represents the benefit of the users, while displaying the many-faceted roles that program slicing plays. In this way users can proceed with program slicing to large systems or unfinished systems.展开更多
This letter proposes a hybrid method for computing dynamic program slicing. The key element is to construct a Coverage-Testing-based Dynamic Dependence Graph (CTDDG),which makes use of both dynamic and static informat...This letter proposes a hybrid method for computing dynamic program slicing. The key element is to construct a Coverage-Testing-based Dynamic Dependence Graph (CTDDG),which makes use of both dynamic and static information to get execution status. The approach overcomes the limitations of previous dynamic slicing methods, which have to redo slicing if slice criterion changes.展开更多
文摘Though Unified Modeling Language (UML) has been widely used in software development, the major problems confronted lie in comprehension and testing. Dependence analysis is an important approach to analyze, understand, test and maintain programs. A new kind of dependence analysis method for UML class diagrams is developed. A set of dependence relations is definedcorresponding to the relations among classes. Thus, the dependence graph of UML class diagram can be constructed from these dependence relations. Based on this model, both slicing and measurement coupling are further given as its two applications.
文摘With the scale of programs becoming increasingly bigger, and the complexity degree higher, how to select program fragments for slicing has become an important research topic. A new type of criterion called interesting index is proposed for selecting parts of procedures or procedure fragments to do program slicing. This new criterion considers not only the subjective aspects in users, namely users' emphasis on the time efficiency, storage capacity or readability,but also the objective aspect in large procedures. It also represents the benefit of the users, while displaying the many-faceted roles that program slicing plays. In this way users can proceed with program slicing to large systems or unfinished systems.
文摘This letter proposes a hybrid method for computing dynamic program slicing. The key element is to construct a Coverage-Testing-based Dynamic Dependence Graph (CTDDG),which makes use of both dynamic and static information to get execution status. The approach overcomes the limitations of previous dynamic slicing methods, which have to redo slicing if slice criterion changes.
