The traditional approach to building and maintaining a system is to build an initial system and then deploy it. Over time, end users identify certain performance `inadequacies` or `inconveniences` in the initial system. These problems are reported back to the system designer, who collects them, regularly rebuilding the system with `improved` performance. A problem is that the design of the system never exactly fits the task environment. The reason for this is in large part due to the pragmatic relations between system, data, and user. In the standard adapted system, information is being gathered about the user which is used to modify the interface. The goal of this research is to develop a methodology for building systems that evolve based on the particulars of a specific data set. A method of system adaptation will be explored that is based on the history of usage of the system for a given task. With each transaction, the system acquires information about the interaction between the user and the system that is subsequently used to adjust the problem-solving behavior of the user/system as it applies to that particular data set. These ideas will be developed as they apply to the construction and evolution of an online manual system. The results of the study will be incorporated into a methodology for building systems that automatically evolve to the specifics of their task environment after they are deployed.