The dynamic nature of knowledge and software evolution and usage present a pervasive challenge to system developers. Discrete attempts to create such systems often lead to a mismatch between system, expectation and a changing reality. The rationale for a Dynamic Feedback Model stems from the need to focus on a continuous and long-term perspective of development and growth in change-intensive environments. This paper makes the case for a reflective learning and knowledge-driven view of software development and presents such a model in a way that accounts for the long-term survival, growth and evolution of software-intensive systems.