By Roberto Verdecchia
Architectural Technical Debt (ATD) is a serious problem in SA, as debts have potentially a high impact on the overall technical debt of a system. For example, sub-optimal decision result in immature architectural documentation. So, how can we better understand ATD, and detect these effectively? What are sources from which we can discover these debt items? Can we do this automatically, and which and how much user input do we need in the discovery?
The proposed method starts from self-admitted ATDs, abstracted code evolution analysis (reverse engineer the architecture from source code, and do evolution analysis) and (manual?) inspection of additional sources. The idea is that these sources can be combined to identify real ATDs. Of course the whole approach needs to be evaluated through experimentation, both from open source and industrial cases.