An Overview of Rendezvous
Current mobile systems do not adequately support the needs of real-time
collaborative applications. Moreover, systems which specifically try to support
real-time collaborative applications do not perform well in mobile environments
where network outages are commonplace and network latencies are extremely high.
This project aims to resolve these issues by developing a software platform
which will expose a novel programming paradigm to these applications which will
allow them to adapt to the harsh network conditions found in mobile
environments.
The interest of the Rendezvous project is firmly focussed on the need for real
time collaborative systems which are able to deal with more serious physical or
virtual difficulties than the average distributed virtual environment. This may
include problems with the networking between devices, meaning that the majority
of their information on one another is wrong a great deal of the time. It may
also include systems for which the widely used rollback mechanism is not
appropriate, for instance where actions are performed in the real world instead
of the virtual world, which is often less forgiving of retracing steps.
This work aims at enabling disparate collaborating entities to gradually
"rendezvous" back to an acceptable shared state over a period of time, and its
implementation will require drawing of valuable background information from a
wide variety of different areas. First and foremost of these is of course the
existing distributed virtual environments, but also the use of reflection and
formal modelling, alongside the study of methods such as those used in control
theory in engineering where real physical systems are gradually brought to the
correct position or value.
