I've been using Squeak to prototype knowledge management and knowledge capture tools. Two open squeak projects are in development. One is an interface to the Lemur information retrieval toolkit. This is being used for experiments in information retrieval (cf. TREC and TDT). The other project is based on adding IR and concept representation capabilities to the SmallBlog project. This project is CokaBlog, a community knowledge mapping blog system. The idea is to build concept maps within a person's blogs and over a community of blogs for knowledge discovery and management. I am currently working with language models for concept representations .