mfor heterogeneous, distributed information systems

Software development methodologies today do not cover sufficiently the matter of smooth evolution of long-lived software infrastructures, as they, for instance and in our main focus, appear in very large, heterogeneous, distributed information systems. These are based, in general, on several underlying autonomous databases and applications. Change and further development, usually, is driven by new user needs and customer requirements from the top, but also by additional offers of information resources at the bottom level. Our notion of 'continuous software engineering' tries to tackle this challenging task in particular aspects. We suggest a methodology based on three principles: (object-oriented) information modeling on the bottom and top layers in parallel, making explicit use of metainformation as a basic concept, and, componentization in a software-architectural sense. Herein, a consistent development of appropriate meta data structures play a central role as well for the analysis steps as also during the design process. Explicit OO models and software components of meta data help in establishing more flexible solutions, for an improved systems evolution.