Abstract: What do you do to make sense of a mass of information on a given topic? Paradoxically, you likely add yet more information to the pile: annotations, underlining, bookmarks, cross-references. We want to build digital information systems for managing such added or super-imposed information and support applications that create and manipulate it. We find that requirements for a superimposed information system can be quite different from those for a traditional database management system: a lightweight implementation, multi-model information structures, "schema-later" data entry, interacting with data that is "outside the box" (controlled by other applications), and support, rather than removal, of redundancy. We report here on SLIMPad, a superimposed application, which was inspired by the "bundling" of information elements from disparate sources we observed in a medical setting. We propose an architecture for superimposed applications and information management. Our prototype components to implement the architecture give flexibility in structuring superimposed information, and also encapsulate addressing, at a sub-document granularity, into a variety of base information sources.