Long-term collaboration and lifelong learning in computer-augmented environments, as examples of non-traditional database applications, are characterized by potential evolutionary change of systems. Computer-augmented environments provide easy and intuitive methods of manipulating computational media by using physical objects. Conventional technological frameworks for computer-augmented environments use static mapping tables for linking physical objects and computational media. However, static mapping tables are not flexible enough to support unpredictable change in long-term activities. Embedded materialized views are introduced to dynamically produce flexible mapping relations for long-term collaboration. View materialization makes derived mapping relations flexible, efficient and highly available. As a first step towards an evolvable computer-augmented environment, the issue of updating embedded materialized views is discussed.