The gap between software engineering and multi-agent systems: Bridging the divide

The agent community to date has done a poor job of convincing business and mainstream software engineers of the value of Multi-Agent Systems (MAS) and agent-oriented software engineering. This is largely our fault. We have tried to offer a completely new way of doing things without providing a sufficient value proposition for throwing away the existing frameworks, methodologies and infrastructures. This position paper examines some of the reasons why we are not succeeding in this task and puts forward some suggestions on how we could do better.