FlowLynx proposes development of an intelligent, web-based legacy code refactoring and conversion system that significantly improves current software re-engineering methodologies. The FlowLynx proposed tools and methods facilitate collaborative human understanding and behavioral algorithm 'mining' directly from sections of unstructured software code. It will then use this 'mined' behavioral information to learn how to detect and extract algorithmic patterns from unstructured code thereby enabling the capture and automation of code understanding, re-engineering and translation knowledge. This effort will demonstrate the feasibility of developing an artificially intelligent de-compiler or 'algorithm detector' that can learn how to refactor low-level code into higher-level algorithmic constructs. Alternatively this system could be used to develop intelligent code optimization tools, generalized de-compilers, plus the detection and repair of damaged code within an 'intelligent operating system' context. Anticipated Benefits/Commercial Applications: Code maintenance, reuse and translation, behavior mining, and de-compiler technology. Aerospace firms have a tremendous amount of legacy code that requires significant reengineering and can benefit from re-factoring and translation into modern languages. This technology will address a critical step in automated re-factoring of primitive code into algorithmic structures that improve code understanding, optimization, language translation and automated analysis.
Keywords: automated, re-factoring, primitive, algorithmic, code, understanding, optimization, translation
