Our reference model is targeted towards the class of systems (or
components of superordinate systems) whose task is to present
information to the user in an effective way. Hereby the
attribute effective means that the particular information needs
of the individual users are best met under a given set of presentation
constraints, such as resource limitations and the user's knowledge and
style preferences. We do not deny that there are applications for
which automated presentation design would be ``oversized''.
Rather, we focus on applications comprising sufficiently complex presentation
tasks, such as medical, educational, or industrial applications which
consitst of large amounts of complex information and users with wide
ranging knowledge, skills, tasks and environments/situations.
Throughout this framework, we always assume a human user being the
addressee of a generated presentation.
In principle, it could be indeed any external entity able to
interpret a presentation. However, if such an
entity is another computer system, this raises the question of whether other
forms of communication are more effective and efficient.
We explicitly choose not to address a more general
class of systems - namely intelligent multimedia dialogue systems.
Although desirable, we consider it too ambitious to handle all at once,
multimedia input analysis and multimedia presentation generation.
Furthermore, excluding full-fledged dialogue systems does not
necessarily mean excluding interactivity completely. Actually, many
interactive systems only provide interaction facilities which do not
require deep analysis of user input. Examples include hypermedia
systems where pointing gestures are associated with commands, and
systems that allow users to input via menus or via a command language.
If a multimedia system does some intelligent
processing but not for the purpose of interpreting input, then the
system may be interesting only because of it's presentation abilities, and
thus we consider it an IMMPS rather than an intelligent dialogue system.
Finally, we stress that the reference model is intended to fit real systems, and that we constantly have been guided towards this goal. However, this does not necessarily mean that the proposed architecture has to be followed in an implementation. The RM simply provides a logical view of the generic tasks which occur in multimedia presentation generation.