To describe the design of HAC we will use one running example. Suppose that the user is working on a project involving the use of fingerprints (as one of the authors had). Information about the project may be found in email with its participants, in notes, articles, source code files, etc. Typically each of these will be stored in a different place, possibly on a different computer (e.g, a laptop or a network file server). The user may also have relevant information from previous projects or from other sources which the user may not even remember. Furthermore, important information can be obtained through a search of remote facilities. HAC allows to combine all relevant material in one semantic directory; let's call it fingerprint. We'll see how to build it, maintain it, and use it later on.