We define formally the data models WP, AWP and AWPS especially designed for the dissemination of textual information by distributed agent systems using communication languages such as KQML and FIPA-ACL. We also define the problems of satisfaction and filtering and point out that these problems are fundamental for the deployment of our models in distributed agent architectures appropriate for information dissemination. One such architecture currently under development in project DIET is sketched in some detail in this paper. Finally, we present algorithms for the problems of satisfaction and filtering, prove the correctness of these algorithms, and calculate their computational complexity.