Hypermedia systems have been demonstrated to support authoring and reading of mostly static information. Few systems address the needs of analysts deriving information from a continuously changing base of information. Those that do, focus on the existing content and use links primarily for navigation and management. An open hypermedia architecture is proposed for a class of analysis systems where the value added by the analyst is through associating data elements. In such systems, links are the primary form of information being managed. The architecture developed provides a framework through which hypermedia analysis systems can be generated with little or no code development. Specifically, the model is shown to apply to the domain of software engineering by mapping the analysis portions of a rapid prototyping lifecycle to a schema defined using the framework. Through the addition of n-ary links and links to links, the architecture provides a closer mapping to the Dexter Hypertext Reference Model than current graph-based models such as the Multimedia Object Retrieval Environment (MORE). Improvements over MORE are also shown in the use of abstraction as a filtering mechanism and through the full involvement of links as being the primary focus of the analysis, query, and filtering functions.