Communication Meaning: Foundations and Directions for Systems Research

The multiagent software research program envisaged putting multiagent abstractions and methodologies at the heart of designing intelligent distributed applications. In particular, with the aim of enabling flexible interactions between agents, it emphasized modeling communication meaning. After three decades of work, the program can claim little broad impact. Has the program been a failure? We think the program has seen remarkable successes, the biggest being the recent work on information protocols, which finally makes it possible to realize the promise of modeling meaning. In this paper, in support of our claim, we set out how information protocols advance the cause of meaning. We then argue that these advances strike against conventional systems wisdom, including in fields such as networks, distributed systems, and programming languages. Finally, we lay out an ambitious research agenda that puts multiagent abstractions at the heart of systems research. Now is a good time to double down on MAS software research.