We present a general model and information server for the digital annotation of printed documents. The resulting annotation framework supports both informal and structured annotations as well as context-dependent services. A demonstrator application for mammography that features both enhanced writing and reading activities is described.