Abstract
The paper focuses on the novel features of the architecture of the "virtual campus" infrastructure we are building at Politecnico di Milano. First, it illustrates the rich underlying conceptual model, which fosters reuse and evolution of learning objects by aggregation and modification. Complex learning objects can be defined and enacted as learning processes by a work-flow engine that manages and distributes learning and working data. Data can be analyzed to reason about users' behavior, audit their work and help them. Autonomous tutoring agents are provided for these tasks. The system provides virtual connection facilities to enable data sharing in untethered contexts.
