The Polimi Virtual Campus platform is an open system for the design, deployment, fruition, and evaluation of reusable learning materials. Its main objectives are:

• To support design, composition, and reuse of Learning Objects (LOs).
• To support fruition of Learning Objects:
  • Traditional and ad-hoc settings.
  • Personal and cooperative usage.
• To support analysis of students' behavior (learning, relational, and normative aspects).
• To analyze power consumption in all envisaged scenario.
• To experiment the usage of our learning platform and of advanced learning objects in our CS courses.

• Development of a conceptual model that extends the IEEE LOM.
• A prototype that integrates various competences (software engineering, database, hypermedia, hardware design).
• Usage of innovative and consolidated technologies from MS and Politecnico di Milano.

Project approach

Reusability of learning materials is key in order to improve the cost-effectiveness of computer aided learning environments. Thus, learning materials in Virtual Campus take the format of "learning objects" (LOs) of arbitrary complexity, typically arranged into hierarchies of smaller granularity LOs, described by a meta-model. A semantically rich description of LOs is useful for understanding their features and for enabling their publishing in a Web context. The richness of LO metadata structure enables their reuse across completely different learning sessions, by improving search capabilities and by providing information supporting intelligent retrieve.
A LO is defined as the smallest unit that can be delivered into the Virtual Campus. It is the smallest unit that can be assigned to a leaner and for which a learner’s position may be traced. Structurally a LO is composed by a metadata part and a content part.
In the Virtual Campus model everything is classified in terms of LOs. The concepts of “class”, “exam”, “course” do not exist in themselves, and are modeled as LOs. Single LO can be aggregated into more complex learning units (in their turn defined in terms of LOs) through the use of “relationships”. The figure below shows LOs and their relationships.

LOs and their relationships

A system for managing LOs should represent them at three different abstraction levels: the reusable-content level, where LOs content is defined in a way to be applicable to multiple contexts; the didactic level, where LOs are seen in the context of a specific didactic use, e.g. a given university curriculum; and the fruition level, where the actual embodiment of a virtual class is defined, by describing the class milestones, the actors participating to the class and their roles relative to LOs, and so on. The design support system should be suitable for the interaction with "content providers" covering three different roles of author, course organizer, and teacher; each role is involved with the relevant level of the specification. Learning processes defined by complex learning objects need to be supported by the system engine. Under this viewpoint the system acts as workflow environment: it manages learning and working data, producing proper outputs that are distributed again to users, in order to support their activities and improve their work efficiency.

Virtual Campus architecure and models

High level architecture

The Virtual Campus platform is composed of an Authoring Environment and a Fruition Environment. The authoring environment allows users to create and manage LOs at the three levels defined in the previous section. The fruition environment provides all mechanisms to exploit LOs for the purpose of providing and obtaining an educational service. The core of the fruition environment is an engine that executes the fruition process defined as a part of LOs. The evaluation environment constantly monitors and collects metrics on the effectiveness of LOs and on the behavior of learners and teachers. All such data provide useful feedbacks that are then used to improve LOs. The figures below provide a view on the architecture.

The three-level model

The Virtual Campus design is expected to maximize re-use (allowing Authors/Organizers to easily modify and assemble LOs) and customisation (allowing Learners to adapt fruition to their personal preferences) of the Learning Objects. In order to achieve these goals, the LO “life-cycle” has been divided into three different phases (or "levels").

Reusable-content Level (RL)

Our strategy is to express CLOs without making use of explicit paths definition. In other words, we’ll use a language that gives Organizers the opportunity to express logic dependencies among LOs (i.e. “knowledge about Limits is required in order to understand Derivatives”, “Analysis3 and Calculus are alternative so Learners can choose either the former or the latter”, etc).
Notice that, using such language, Organizers don’t impose an a priori ordering to LO fruition, limiting themselves to the definition of the constraints a given path must respect. Therefore, Organizers can focus on didactical relationships among LOs (what pieces of knowledge a student must learn in order to understand a given LO, whether diverse LOs give the same knowledge, whether a piece of knowledge is compulsory or represents an optional didactical integration, etc).

Didactical Level (DL)

In DL, we make use of a “workflow-like” representation (using UML Activity Diagrams). The system translates DL language into a workflow and give Teachers the opportunity to customize the resulting workflow, by removing/modifying paths. Since we use a workflow representation, Teachers can focus on customizing more precisely the fruition of their courses. Starting from the same RL-LO (i.e. "Analysis"), Teachers can derive multiple DL-LO (i.e. "Analysis for mathematics" and "Analysis for physics"); each one represents a sort of "specialization" of the RL-LO.

Fruition Level (FL)

Starting from DL, further Preconditions and Postconditions are added and the system is able to generate a particular description (FL) used as input of a workflow application. The LO is now "published" and the workflow application will control the fruition of the CLO, presenting to each Learner the LOs she is allowed to use. Starting from the same DL-LO, Teachers can publish multiple FL-LO (i.e. "Analysis for physics, 1st semester 2003" and "Analisys for physics, 2nd semester 2003"); each one represents a sort of "instantiation" of the DL-LO.

Virtual Campus three-level model

Virtual Campus metadata (VC-LOM)

Metadata is data about data, descriptive information about resources for the purpose of finding, managing, and using them more effectively. We use this system of labels to describe a resouce or object's characteristics and its objectives. The purpose of metadata is to provide a common means of describing things (electronically) so that we can define, search, and find learning objects (however they are defined).
The goal of the Virtual Campus information model (VC-LOM, Virtual Campus - Learning Object Metadata) is to specify of the LOs' structure in terms of metadata and contents (references to raw data). The VC-LOM is based on an extended version of the IEEE LOM. The figure below shows VC-LOM main components, data elements describe a learning object and are grouped into categories. The base scheme consists of nine such categories.

VC-LOM main componenents

We focus especially on Relationships beetwen LOs, Precondition/Postconditions and Interaction modalities.

Relationships

Semantics of IEEE LOM Relation metadata is not so clear. We introduced a new vocabulary for Relation metadata and defined precise semantics for each relationship. Then, we used relationships in order to express didactical dependencies beetwen LOs (i.e. "In order to undertand Derivatives, a student must study Limits" is represented as "Limits IsRequiresBy Derivatives") that represents constraints to the fruition of LOs. We use Relation metadata to store our RL language. Due to the precise semantics we defined for each relationships, a translation process can generate a workflow description (our DL language). The generated workflow will respect the constraints introduced using the relationships.

Preconditions/Postconditions

Preconditions and Postconditions have been added to express particular constraints that regulate LO access. As an example, we consider time constraints (start time and end time of a LO representing a live lesson), administrative properties of Learner, educational requirements a learner should possess in order to exploit successfully the LO (however, these are just suggestions: the system doesn't consider them during the translation process), etc.

Interaction modalities

VC-LOM supports several "interaction modalities" in order to make use of diverse LO contents and allows cooperation among Learners and between Learners and Educators. In particular:

• LO fruition can either be individual or require a group participation. If a group is required, LO metadata must specify the related cardinality.
• LO collective fruition can either be "synchronous" (all participants to a group must be on-lime at the same time) or "asynchronous".
• LO fruition can either be "auditable" by other Learners or "not auditable".
• LO fruition can require the presence of one or more "guides" (supervisors). If supervisors are required, LO metadata must specify supervisors’ rights and group cardinality.