Software Engineering

摘要

Over the years, visualization tools have evolved to meet the increasing demands of the software industry. Yet, as software projects grow to monumental size and complexity, the need for a new generation of tools is more pressing than ever. Most of the existing construction tools offer nothing more but sophisticated graphical editors for visualization. While these tools help in the analysis and design of small to medium software projects, they are useless for large, complex applications. In this case, it is imperative to complement them with powerful layout features, such as the ones developed by the graph drawing community (see section 2.3). In addition, very few construction tools offer dynamic visualizations, i.e., animation and simulation. It is apparent that more research is needed in this area. Finally, even though some visual programming environments are available on the market, these tools are still not widely used by the software engineering community. It may be interesting to assess these tools and determine the features that will make them more appealing to software engineers. With respect to analysis tools, it seems that most of the work has concentrated on the development of elaborate information extraction methods. Most of the analysis tools rely on general purpose graph layout tools to visualize information. Because these tools are generic, they do not take into account the drawing conventions of accepted software engineering notations (e.g., UML, data flow diagrams, etc). Additionally, these layout tools do not scale. In re-engineering, software artifacts may necessitate the visualization of graphs consisting of tens of thousands of nodes. Hence, in addition to improving the existing layout algorithms, it is necessary to develop new navigation mechanisms to browse through complex representations of code. Or better, it may be the time to investigate new visualization techniques. This chapter starts with a paper by Rainer Koschke from the University of Stuttgart discussing visualization tools for reverse engineering. The author describes Bauhaus, a system which recovers architectural designs from C programs. The visualization of the recovered information is produced by the generic graph editor of Rigi, and the general purpose graph layout tools Graphed and Graphlet. Wim De Pauw et al. from IBM Watson Research Center describe Jinsight, a tool that visualizes the run-time behavior of Java programs. Jinsight allows the user to explore program execution through several linked views (e.g., histogram view, reference pattern view, execution view, etc.). Katherina Mehner from the University of Paderborn discusses a visualization and debugging environment for concurrent Java programs. The UML-based visualization is achieved by integrating the tool with Together. Rainer Oechsle and Thomas Schmitt from the Universities of Trier and Saarbruecken describe JAVAVIS, a system used to visualize the dynamic behavior of sequential JAVA programs. The tool displays the object and sequence diagrams of a running program. Thomas Zimmermann and Andreas Zeller from the Univerity of Passau discuss the visualization of memory graphs. A memory graph is a construct that captures the state of a running program at specific times during its execution. This paper discusses an analysis and extraction method. The visualization is produced by the generic graph layout tool Graphviz.