This paper presents an approach and tool to automatically instrument dynamic web applications using source transformation technology, and to reverse engineer a UML 2.1 sequence diagram from the execution traces generated by the resulting instrumentation. The result can be directly imported and visualized in a UML toolset such as Rational Software Architect. Our approach dynamically filters traces to reduce redundant information that may complicate program understanding. While our current implementation works on PHP-based applications, the framework is easily extended to other scripting languages in plug-andplay fashion. In addition to supporting web application understanding, our tool is being used to recover traces from dynamic web applications in support of web application security analysis and testing. We demonstrate our method on the analysis of the popular internet bulletin board system PhpBB 2.0. 1

Abstract—Building comprehensive test suites for web applications poses new challenges in software testing. Coverage criteria used for traditional systems to assess the quality of test cases are simply not sufficient for complex dynamic applications. As a result, faults in web applications can often be traced to insufficient testing coverage of the complex interactions between the components. This paper presents a new set of coverage criteria for web applications, based on page access, use of server variables, and interactions with the database. Following an instrumentation transformation to insert dynamic tracking of these aspects, a static analysis is used to automatically create a coverage database by extracting and executing only the instrumentation statements of the program. The database is then updated dynamically during execution by the instrumentation calls themselves. We demonstrate the usefulness of our coverage criteria and the precision of our approach on the analysis of the popular internet bulletin board

TXL is a functional programming language specifically designed for expressing source transformation tasks. Originally designed for the rapid prototyping of modest syntactic enhancements, in recent years it has been extensively used in large scale source code analysis and reengineering applications that are much more challenging. As a result, many common programming techniques needed in these larger scale applications are difficult or impossible to express in TXL. Examples include multi-way decisions, generic rules and functions, polymorphism and information hiding. In this paper we introduce ETXL, an experimental extension of TXL which includes convenient features designed to address these issues. Designed to be a compatible variant that remains faithful to the original TXL syntax and semantics, ETXL has itself been prototyped as a source transformation to original TXL.

One aspect of autonomic computing is the ability to identify, separate and automatically tune parameters related to performance, security, robustness and other properties of a software system. Often the response to events a#ecting these properties consists of adjusting tuneable system parameters such as table sizes, timeout limits, restart checks and so on. In many ways these tuneable parameters correspond to the switches and potentiometers on the control panel of many hardware devices. While modern software systems designed for autonomic control may make these parameters easily accessible, in legacy systems they are often scattered or deeply hidden in the software source.

Abstract. Enrichment of text documents with semantic metadata reflecting their meaning facilitates document organization, indexing and retrieval. However, most web data remain unstructured because of the difficulty and the cost of manually annotating text. In this work, we present Cerno, a framework for semi-automatic semantic annotation of textual documents according to a domain-specific semantic model. The proposed framework is founded on lightweight techniques and tools intended for legacy code analysis and markup. To illustrate the feasibility of our proposal, we report experimental results of its application to two different domains. These results suggest that light-weight semi-automatic techniques for semantic annotation are feasible, require limited human effort for adaptation to a new domain, and demonstrate markup quality comparable with state-of-the-art methods.

ter verkrijging van de graad Doctor aan