Documenting Early Architectural Assumptions in Scenario-Based Requirements

In scenario-based requirement elicitation techniques such as quality attribute scenario elicitation and use case engineering, the requirements engineer is typically forced to make some implicit early architectural assumptions. These architectural assumptions represent initial architectural elements such as supposed building blocks of the envisaged system. Such implicitly specified assumptions are prone to ambiguity, vagueness, duplication, and contradiction. Furthermore, they are typically scattered across and tangled within the different scenario-based requirements. This lack of modularity hinders navigability of the requirement body as a whole. This paper discusses the need to explicitly document otherwise implicit architectural assumptions. Such an explicit intermediary between quality attribute scenarios and use cases enables the derivation and exploration of interrelations between these different requirements. This is essential to lower the mental effort required to navigate these models and facilitates a number of essential activities in the early development phases such as the selection of candidate drivers in attribute-driven design, architectural trade-off analysis and architectural change impact analysis.