A Decision Framework on Refactoring Architectural Technical Debt: Paying Back in Modularity- An Industrial Case Study

Abstract

Technical debt refers to sub-optimal solutions during software development where there is a trade-o between short-term and long-term goals. Lately there has been a few studies which identi es technical debt, however most of the do not estimate the interest which is associated with the identi ed debt. Knowing how much interest is being paid allows the developers to make informed decisions of what will bene t the development. One example is knowing if a onetime cost of a repaying the debt outweighs the cost of paying the interest of that debt. This would mean that the repayment can be seen as an investment for future development. This thesis aims to develop a decision framework that can be used when deciding if part of a component would bene t from being modularized into a new component or framework to repay a debt. To accomplish this, the study developed two methods that are used by the decision framework. The rst method is to nd out if the analysed part of the components would bene t from such a modularization. The second method estimates how much e ort can be saved by doing a modularization. It was found that for the rst method, a measurement system which analysed the component's source code was a good approach in deciding if a modularization would be bene cial. For the second method an approach which used data regarding the current e ort distribution to estimate the e ort saved by modularizing was chosen. The result of combining the two methods was found to be an adequate decision framework which provides useful information in the decision if to modularize part of a component or not.