In a new method of representation and use of knowledge, the separation of knowledge from data is maximized. The knowledge and data are physically separated both in the description and in the representation: the knowledge is concentrated in the information structure, while the representation and the data processing are located in terminal algorithms. These two components are represented by two different subspaces of the common space for problem solving. The separation of spaces and mechanisms of data and knowledge processing allows the user to apply the optimal hardware interpretation tools, while preserving the unity of the two-pronged problem solving process.