Release managers often face a dilemma about the quality of software under delivery before a release. The presence of run-time errors such as memory leaks, buffer overflows, and deadlocks affects quality attributes such as efficiency, security, and reliability. Such errors are detected using dynamic analysis methods in practice. However, the dynamic analysis methods employed in practice are by and large ad hoc. It is essential to use dynamic analysis focusing on finding the right set of run-time errors in a software component that have the maximum impact on quality. There exists a need to identify quality attributes such as reliability, efficiency, and security that are important for a software component, or, a system. In this paper, a quality model driven dynamic analysis methodology is proposed. Various run-time errors that can arise during the execution of programs written in a language such as C++ are mapped to the respective quality attributes, thereby forming a basis for run-time error classification. Our experiences in the application of dynamic analysis on real projects are reported. The methodology reports the error findings mapped to the quality attributes along with their distributions. The reported findings help management understand quality problems and take appropriate corrective action.

This tutorial provides an introduction to the problems faced in software evolution, and why we need to be concerned about refactoring design and code on a continuous basis. Such refactoring helps ensure that software is more easily changeable, and the costs of development and maintenance are reduced. This tutorial addresses (a) refactoring techniques that a developer can employ during design and coding and (b) static analysis techniques that detect opportunities for refactoring.

A compile-time analysis technique is developed to derive the probability with which a user-defined function or a supercombinator requires each one of its arguments. This provides a basis for identifying useful speculative parallelism in a program. The performance of speculative evaluation is compared with that of lazy evaluation, and the necessary conditions under which speculative evaluation performs better are identified.>