To comprehend an object-oriented software, visualizing object interactions performed as a UML sequence diagram is a promising technique for helping developers. However such UML sequence diagrams are typically too huge to comprehend behavior. In this paper, we propose a novel technique to generate abstracted sequence diagrams based on the information of applied GoF design patterns in a source code. We discuss effectiveness of our approach through a case study applying our implemented tool to an open source software.

We propose an XML C source code representation to support developing CASE tools. Since source code is a main artifact of software development, most CASE tools have some features related to source code editor, static analyzer, profiler, etc. To develop such tools, detailed information related to source code is needed. However, it is quite difficult to reuse program analysis features because they do not have common interfaces even for parsing and data/control-flow analysis that are most common features for such CASE tools. To address this issue, we focus on XML as an intermediate representation for source code information. Existing XML representations only represent structure of syntax trees and lack some important information required for CASE tools. We propose two models for representing source code, one is for intra-file information, which consists of syntax structure, flow, and type information, the other is for inter-file relation, which is cross-reference information. We also introduce CASE tools with our representation and demonstrate the efficacy in CASE tool development. To evaluate the efficacy, we show that a coding rule checker and a cross-referencer can be easily implemented using common XML processing libraries such as XSLT and XPath.

Software metrics is one of promise technique to capture the size and quality of products, development process in order to assess a software development. Many software metrics based on various aspects of a product and/or a process have been proposed. There is some research which discuss the relation between software metrics and faults to use these metrics as the indicator of quality. Most of these software metrics are based on structural features of products or process information related to explicit fault. In this paper, we focus on latent faults detected by static analysis techniques. The coding checker is widely used to find coding standards violations which are strongly relating to latent faults. In this paper, we propose new software metrics based on coding standards violations to capture latent faults in a development. We analyze two open source projects by using proposed metrics and discuss the effectiveness.

Software visualization with sequence diagrams is one of the promising techniques aimed at helping developers comprehend the behavior of object-oriented systems effectively. However, it is still difficult to understand this behavior, because the size of automatically generated sequence diagrams tends to be beyond the developer's capacity. In this paper, we propose a sequence diagram slicing method, which is an extension of our previous method based on a dynamic slicing technique using static information. Our proposed method is capable of accurate slice calculation based on high-precision data dependency and can support various programs, including exceptions and multithreading. In addition, our proposed new tool performs slice calculations on the Eclipse platform and we demonstrate the applicability of this method by applying the tool to two Java programs as case studies. The results confirm the effectiveness of our proposed method for understanding the behavior of object-oriented systems.

We propose a compositional analysis method for synchronization flow in AIBO OPEN-R programs based on communicating processes. Concurrent objects of AIBO programs with the OPEN-RAPI are synchronized by two types of signals: ready and notify. Focusing on these signals, we describe abstract behavior of AIBO programs in the pi-calculus preserving the source code structure. We model-check deadlock freeness and interactions order of AIBO programs based on this abstract behavior. Since a primary issue in model-checking is the state space explosion in the behavioral model, we present a decomposing method to reduce the combination of states of concurrent objects. Since our translation to the pi-calculus preserves the syntactical structure of source code, when a counter-example is pointed out, our method enables not only to detect the violation of property of the whole system, but also to point out which component may cause the violation. We developed a prototype translator from AIBO OPEN-R programs to abstract description in the pi-calculus. We show that an application of our decomposing method enables to practically model- check properties by existing tools in two examples of AIBO programs.

Computer-aided requirements definition methods/environment are presented. A composite requirements model (requirements frame) is proposed, and several supporting methods of software requirements definition based on the model are developed. They are: analysis of requirements; describing requirements; verification of an SRS (software requirements specification); SRS execution; and software preliminary design. With these methods, it is possible to obtain an SRS of good quality (correctness, testability, traceability, feasibility, and usability). The authors have been developing an environment of requirements definition, named CARD, by using these methods. They present some examples of experimental uses of CARD.>