Today, more and more people in knowledge communities, like research laboratories, use shared file servers to store and share their information. People in such communities often work together and their files stored in a file server have relationships with each other. Information on the relationships is usually exchanged offline and used implicitly to facilitate the management and sharing of the files. This paper proposes new functions to manage and use the relationships to make various views on the file servers. The functions provide a high-level support and are compatible with the operational framework of existing file systems.