Other agents have been designed to monitor user tasks and suggest resources to users. Some such agents are able to make much more specific suggestions than WordSieve is able, usually at the expense of needing a specific domain model, such as pre-created categories. Lumiere is an example of such an agent apporach . Another example focused on information retrieval is found in Chen . Other agents do not require a specific domain model and suggest documents based on vector comparisons with current document content [7,2,8]. Some agents use explicit feedback to enhance contextual information [4,1]. One of the advantages of WordSieve is that it does not require user feedback. However, if desired, it could be modified to take utilize such information.