One of the most important things an architect can do is reflection. That is, examine
systems, organizations, people and ask “What alternatives were considered and why
was that particular decision made?” Thinking about the response gives an architect
insight into the motivations and decision processes that others have used and this, in
turn, should help the architect make better decisions in the future. A pre-requisite for
doing this type of reflection is that the decisions and alternatives are made explicit.
One venue that gives an architect an opportunity to do this type of reflection is during
an architectural evaluation.Another venue is from a book such as this. This book
lays out the design process used in building a collection of multi-agent systems.
In addition to providing a case study of a design process and the rationale for
the design decisions, the topic of the book also is of great interest. Systems of
the future will increasingly have the characteristics of the autonomous systems
described here: they are simultaneously becoming more interconnected and more
autonomous. Think of your smart phone that is mostly connected but can operate
many functions while it is disconnected. These systems of the future will also
increasingly operate without central control. Again, the telephone system and how
cellular communication is managed provides a good example of this phenomenon.
Problems of connectivity raise issues of a node discovering that it is disconnected,
other nodes discovering that a particular node is disconnected, how the
node operates while it is disconnected, and reconnecting the node. The case study
provides solutions to this problem in the context of autonomous vehicles within a
limited geographic area. The essence of the solution provided—define the concept
of a neighborhood for a node and treat neighboring nodes in a different fashion from
other nodes—seems like it is more general than the particular application in the case
study but that is still to be determined.