Experiences of Using Extreme Programming to Support a Legacy Information System Migration Project

In the recent years, the software engineering community has been showing a growing interest in agile development methodologies. These have emerged out of the need for faster, more flexible and efficient processes for software development. Currently the best known agile method is extreme programming (XP). XP addresses issues concerning changing requirements and their cost by simplifying management tasks and project documentation. XP uses an iterative and incremental software process executed in relatively short cycles. Traditionally this type of approach would yield an increased management overhead because the management activities related to the ending and starting of iteration have to be executed for every iteration, while in the XP process these activities are minimized. Currently, there are lots of experience reports available on applying the XP method to different types of software projects. However, there exist very few experience reports of XP projects where the database is the most central issue of the project. Thus, the objective of this paper is to report and examine experiences and empirical data while using XP practices to support a legacy information system migration project. The work was performed during 4 months, with a total effort of 1056 hours. The project followed the agile methodology principles by applying a series of XP practices (e.g. planning game and small releases). Two previous migration attempts had failed. These efforts had followed the traditional waterfall approach, finishing already in the design phase, producing mainly design documents, but no description of the technique for transferring the data. In addition, the design documents produced were highly complicated and did not completely correspond to customers' needs. This paper attempts to understand the reasons for these failures, while also compiling a list of lessons learned. This study brings empirical data supporting the idea that XP practices and documentation for need are applicable not only to pure development projects (where source code is the most important result), but also for other types of software projects, where the main activities may be design (not product) oriented. The project was an engineering success, managing to find solutions to migrating data to the format required by the new information system. The migration itself is planned to be carried out in the near future.