Chinese Journal of Computers   Full Text
Title	Component Integration Test Based on a Metadata Model of Coupling Testing Information
Authors	MA Liang-Li GUO Fu-Liang LI Yong-Jie
Address	(Department of Computer Engineering, Naval University of Engineering, Wuhan 430033)
Year	2007
Issue	No.10(1705—1712)
Abstract &   Background	Abstract This paper introduces a method to increase testability and understandability of component based on definitions and uses information of interface variables about component provided by component developers. Then formal definitions of component coupling testing criterion, definition-use attributes (DU-As) and observation-point values (OP-Vs) are given. Based on these, a definition-use table(DU-Table) is introduced, which includes DU-As and OP-Vs item. Then a framework of testable component based on above DU-Table is given. Above methods are applied to the component developed by themselves before, and related test cases are generated. Moreover above method is compared with Orso method and Kan method using same example, presenting the comparison results. The results illustrate the validity of above method, effectively generating test cases and finding more mutants. keywords component; component system; component-based software integration testing; component coupling testing; component metadata background In recently years, component-based software technology has been increasingly considered as necessary for the vastly more complex software. The main motivation is the possibility to significantly reduce development costs and time yet satisfying quality requirements. But in many cases, the drawbacks of component-based software technologies arise because of the lack of information about provided components. So software development organizations are facing unrivaled challenges. The simultaneous use of many software package developed in-house as well as (often customized) off-the-shelf software has further intensified the existing integration problems of application systems. So how to assure component-based software systems (CBS) quality is made more challenging. It is well known that component-based software testing is one of the most effective techniques assure the quality of component-based systems. Generally, testing of COTS software is black-box because users do not have access to the source code to analyze the internal implementation. How to improve the poor testability of components is one of the issues and challenges techniques that use component testing. To increase the testability of component, it is essential for component users to know the interior details about the component under test. Several years ago Orso presented an idea using metadata describing information about component to solve the problem. Component developers provided the detail of component without revealing source code. Existing component standards, including DCOM, Enterprise JavaBeans and Microsoft .NET already supplied some additional information about a component through the use of metadata that are packaged with the component. But the metadata available so far are typically limited to information that is useful for compile-time and run-time type-checking and for design-time customization There are some related researches addressing to increase the testability of component. In this paper, the authors propose a new analysis method based on some definition and use information extracted from source code by component developers. The information is represented in the form of table, which is encapsulated with the component as component metadata. And the information is used to increase component testability. Based on analyzing the testability of component, a method how to construct a testable component, including related data flow analysis is given.