A multiagent system, conceptually, is often explained in terms of a collection of agents, each of which is loosely connected with each other via Internet/Intranet and possesses the capabilities claimed such as autonomy, adaptation, and cooperation using speech acts like KQML. The agent-view of software is indeed very powerful and will keep attracting more and more people into this fantasy. But years from now, an alternative view of multiagent-based software must be developed in order to raise the quality, productivity, and scalability as well as to reduce the complexity and delivery time. Accordingly, the building environments in support of such an alternative view should be developed. This papers first illustrates the new view of next generation multiagent-based software, product-line architectures, and then presents a building environment towards such an alternative view of multiagent-based software.
Lecture Notes in Computer Science Volume 1599,1-17