Software Re-Engineering of the Human Factors Analysis and Classification System b1s (Maintenance extension) using object oriented methods in a Microsoft Environment

摘要

The purpose of this research is to technically evaluate, refine, and expand two existing aircraft safety management information systems (one military and one civilian). The systems are used in the data collection, organization, query, analysis, and reporting of maintenance errors that contribute to Aviation mishaps, equipment damage, and personnel injury. Both programs implement the Human Factors Analysis and Classification System (HFACS) taxonomy model developed by the Naval Safety Center (NSC) to capture aircrew errors in Naval Aviation mishaps. The goal of this taxonomy is to identify areas for potential intervention by fully describing factors that are precursors to aircraft accidents. Requirements outlined by Dr. John K. Schmidt of the Naval Safety Center, in conjunction with funding by the National Aeronautics and Space Administration, require that the system utilize a Microsoft Access based implementation. This research focuses on meticulous software engineering to investigate the feasibility of adapting the current "structured" systems to Microsoft-based object oriented architectures ensuring future scalability and increased potential for code-reuse. Primary research questions investigated in this thesis include: 1) How can a Microsoft Access-based implementation provide multi-user access to the same database in a client-server environment while ensuring the ability to scale to a large number (potentially thousands) of users? 2) How can the linguistic discontinuity associated with object-oriented concepts and non-object oriented, flat relational databases be overcome when limited by the requirement for a Microsoft Access based solution? This problem is commonly called "impedence Mismatch". 3) The current military and civilian systems provide similar functionality, but use different database schema. How can object oriented methods be implemented to provide a common interface to both types of data? 4) How should database schema be changed to provide the best performance, scalability, and opportunity for code re-use? 5) In the past, Microsoft has deployed new versions of Microsoft Access and Visual Basic that were not (fully) backwards compatible with previous versions. This caused great discontent among users of applications designed to run under the older versions of these programs. How can our system(s) be designed to isolate them from problems associated with new versions of Microsoft products? Specifically, the pending release of Microsoft Office 2002, the new SQL Server 2000 database engine, and Microsoft Visual Basic.NET. This thesis describes our use of the Spiral Development Model to create a Microsoft Based solution for the Aviation Safety School requirements. We hypothesize that the prototype produced as a part of our research will greatly enhance current HFACS-capabilities and provide the means to weather further changes in requirements and application platforms.