Thank you for this informative post. You clearly outlined the role of agent communication languages (ACLs) such as KQML in shaping early multi-agent systems. Highlighting its historical significance while acknowledging its limitations, such as a lack of formal semantics and interoperability issues, is vital in understanding its reduced use today (Cohen and Levesque, 1995; De Ridder, 2025).
Your comparison between ACLs and modern programming languages is relevant. Java and Python have become the dominant choices in agent development, supported by frameworks such as Java Agent DEvelopment Framework (JADE) and Python Agent DEvelopment Framework (PADE). These frameworks provide integration with web technologies, standard libraries, and robust communication protocols, which make them more adaptable for real-world deployment.
While ACLs like KQML laid the foundation for agent interaction, the shift toward using general-purpose languages reflects the need for practicality and ease of use. However, it’s worth noting that ACLs still offer value in systems that require high-level semantic communication, such as in negotiation or coordination tasks (Mayfield, et al., 1996). These capabilities may not be easily replicated using standard method calls or REST-based messaging.
Overall, your post presents a balanced view of past and present communication approaches in agent systems. It highlights how agent-oriented programming continues to evolve by drawing on both foundational concepts and modern development tools.
References
Cohen, P.R. and Levesque, H.J. (1995) ‘Communicative actions for artificial agents’, Proceedings of the International Conference on Multiagent Systems. San Francisco, California, 12-14 June. Cambridge, Massachusetts: The MIT Press. 65-72.
De Ridder, A. (2025) Comparing Agent Communication Languages and Protocols: Choosing the Right Framework for Multi-Agent Systems. Available at: https://smythos.com/ai-agents/ai-agent-development/agent-communication-languages-and-protocols-comparison/ (Accessed: 18 April 2025).
Mayfield, J., Labrou, Y. and Finin, T. (1996) ‘Evaluation of KQML as an agent communication language’, in Intelligent Agents II Agent Theories, Architectures, and Languages. Berlin, Heidelberg: Springer Berlin Heidelberg (Lecture notes in computer science), pp. 347–360.