Self e-Learning Networks will manage RDF descriptions of learning objects and
users, which are likely to be stored either as serialised XML or in
graph/triple form.  The aim of this deliverable is to describe technologies
for implementing the reactive functionality of a Self e-Learning Network,
including facilities such as detecting changes in learning object or user
descriptions and notifying users who have subscribed to be notified of such
changes, and automatically propagating changes in the description of one
resource to other, related, resources.

This deliverable starts by summarising the aspects of the user requirements
that will be provided by SeLeNe's reactive functionality, and discussing why
event-condition-action (ECA) rules are a natural candidate for implementing
such functionality.  An ECA rule language over XML data is then described, and
a prototype implementation of the language that allows the definition and
processing of such ECA rules over a centralised XML repository.  Some
directions of further research for this language and its implementation are
identified.

An second ECA rule language, this time operating on a graph/triple
representation of RDF, is also presented, together with some example rules
over SeLeNe descriptions relating to learning objects and users.  A
distributed architecture for the deployment of such ECA rules, that can be
used in cooperation with other SeLeNe services to provide reactive
functionality, is described next.  A peer-to-peer deployment scenario of this
architecture is then addressed in more detail, including an indexing technique
for RDF descriptions throughout the P2P network, a discussion of the way ECA
rules are registered and executed by the system, and a discussion of how the
system adapts to changes in the peers' RDF repositories.  Some directions of
further research for the language and the architecture are finally identified.