The Impact of Distributed Object Technology on Reengineering - Panel Session Introduction

Reengineering is commonly accepted to mean any activity that improves one’s understanding of a softwareintensive system, or that improves the system itself. Traditionally, the approach taken to reengineering has been to understand the system’s structure and to extract its essential functionality so that it can be transformed into a more evolvable system over the long term. But distributed object technology is dramatically changing the nature and economics of legacy system reengineering in two important ways. The first impact is the new approach to reengineering that distributed object technology offers. Traditional reengineering is based on “deep” program understanding and reverse engineering. The cost/benefit ratio of this approach is staying the same in the face of new technologies such as CORBA, Java, and the Web. However, the benefits of “shallow” interface understanding and component wrapping using these distributed object technologies is rising rapidly relative to the replacement cost. As a result, the economic balance is changing from traditional transformation-based reengineering to wrapping-based reengineering. This may have significant impact on many organizations struggling with modernizing their systems in light of the so-called “Year 2000” problem. The second impact is on the reengineering of systems that are built using distributed objects. Current research is primarily focused on using static analysis techniques to reengineer monolithic systems. This approach may not be successful when applied to distributed systems built from off-the-shelf components that are essentially black boxes. New analysis techniques may be required to reengineer such systems. For example, binary reverse engineering, interface behavior probing, and protocol analysis may be more useful in understanding the nature of such a system. 2 Panel format