The information necessary for materials research and development is complicated in its format and is huge in amount. Therefore, neither conventional database management systems nor knowledge bases in artificial intelligence systems can accept information in a way that is sufficiently flexible as required. Another crucial requirement of information systems is to provide functions to process the meaning of information such as analogical reasoning, induction, and abduction. These are essential operational elements in open world problemsolving systems. In other words, next generation systems should have the capability to generate new information as candidate solutions. Accordingly, data in these systems should contain information representing relationships among data that are comprehensive in the domain of concern. Self-organized information is devised to fulfill the requirements and is shown as an example.