Learning-by-doing software: When to choose advise

While there is growing awareness of the value of developing computer-based learning-by-doing environments, there is still a need to analyze the strengths and weaknesses of particular architectures. This paper examines some of the common design issues faced when building learning-by-doing environments and analyzes the relationship between those design issues and the key features of learning architectures, especially an architecture we are currently developing called Advise. We also discuss the circumstances under which Advise is appropriate and when other architectures would be a better choice.