Existence of thresholds in learning has been long recognized. Their nature has been well characterized, and it is recognized that they need to be treated differently from other core concepts.
Helping learners cross thresholds has been identified as one of the challenges of course and curriculum design, and best practices for integrating thresholds into the classroom via active learning have been identified. Threshold concepts have been recognized to exist in computing education as well. In this paper we examine two difficult concepts in software development: model selection and substitutability. By treating them as threshold concepts and applying the recommended practices, we have succeeded in spiraling them into a course, with suitable scaffolds, over a period of several weeks. In the course of doing this, we developed examples that enable students to find relevance through a broader view of substitutability, and also developed a novel approach to writing requirements that helped students understand why FSM models are preferred in some situations.