Abstract:

Modern learning systems must consider at least two factors: contents and access modalities. As a matter of fact, not only has the nature of e-learning material changed deeply, but so have devices and connection technologies for their fruition.
In this paper, these two features and their relationships are analysed and an architecture is proposed which aims at reaching a good compromise between quality of learning and studying conditions.
As for contents, a wide range of heterogeneous data and activities are at disposal: videoconferencing, remote laboratory, chat for cooperative work, textual exercises, tests, etc. [1]-[5]. Some observations:

1. the data format of such material is intrinsically heterogeneous, ranging from plain documents to multimedia files;
2. some activities need to be synchronised, other do not. For instance, an interactive remote laboratory experience with fellow students must obey precise timetables. On the contrary, some exercises can be made with a certain degree of independence from other people and other tasks;
3. there can be many ways to achieve the same target, each corresponding to different data formats and technologies. For instance, a lesson can be taken using videoconference or studying the course’s material (slides, etc.).

As for access modalities and m-learning facilities [6]-[8], the following scenarios are considered: UMTS (e.g.: for PDAs), DSL (e.g.: at home), WiFi (e.g.: within study lounges and libraries), wired/fiber (e.g.: within laboratories).

All these features make e/m-learning flexible, but they also need a particularly robust organisation of contents and management of educational paths. Some rules are consequently stated, which are the counterparts of (1)-(3) with respect to connection technologies.
Among all the activities that a student can undertake, only those can be performed that obey to the following bonds:

a. a connection technology can not always be chosen freely: there are tasks that require a minimal connection speed. For instance, a remote laboratory experiment in cooperative modality can only be fulfilled using at least a good DSL technology. On the contrary, simple exercises can be made using a PDA;
b. only those activities must be made available which are: (i) compatible with the student’s current technology, (ii) compatible with the e-learning plan, (iii) asynchronous or (iv) can be synchronised with other tasks;
c. if a task can be faced by means of different technologies, data of different nature must be at disposal (videoconference vs slides).

All such factors considered, the proposed architecture consists of three-layers:
Lower layer: the data repository stores a hierarchy of subjects and tasks. Each task is associated to: data, data format, minimal connection technology required, possible alternative contents of different format, prerequisites.
Middle layer: it represents each user’s possible paths and checks their feasibility and compatibility. Due to (1-3) and (a-c), such paths’ are time-labelled graphs [9]-[10], whose nodes’ schema is the same as in the lower layer. Compatibility controls are defined by the edges themselves: there is an edge from Ni to Nj at time t if Ni e Nj obey (a-c). This can be achieved through a modified timed Petri Net [11]-[12].
External layer: it makes the user and the system communicate, also by means of different devices. This can be made through XML conversions.

Keywords:

m-learning, heterogeneous/synchronous/asynchronous activities, graph-shaped learning paths, petri.