(A) maintenance support expert system for legacy systems : UNIK-METASOFT

Abstract

Most legacy systems are currently doing important organization jobs in many industries and societies. At the same time, these legacy systems are easily confronted with maintenance crisis, as predicted. Even though there has been a lot of research done in the hopes of solving this problem, because of poor system documents and lack of man power, it is still very difficult to resolve maintenance problem found in real world legacy systems. So, this dissertation proposes an object oriented methodology that can overcome such circumstances in legacy systems, and we have developed a tool, UNIK-METASOFT to systematically support our methodology. Our methodology can be divided into two phases, System Information Construction Phase, which implies both understanding the system, and Restoration and Maintenance Support Phase. System Information Construction Phase can be divided into two substages: understanding and restoration. We also devided Maintenance Support Phase into two substages, Maintenance Requirement Query and Evolution Phase. The first step of applying our methodology to a legacy systems is the System Information Construction Phase. This phase generates a system information base and now, based on the initial acquired information, our methodology can enter into the second phase: Maintenance Requirement Query Phase. The first encountered subphase of the maintenance support phase is the Maintenance Support Phase. This phrase requires input from the maintenance requirement system, and imformation from the current system. It will then return to the resulting region to be maintained. After maintenance work, the last phase, Evolution Phase is invoked to support reporting. In addition it will adapt the system information to be consistent with the new state of the system. Moreover, the proposed methodology and UNIK-METASOFT are applied to the real world problem of a legacy system from a leading company in Korea. It appears to be very successful. The application empiric...