Relational Pattern Analysis Tool for Legacy Conversion

This project develops a technology to automate the interpretation and understanding of complex trees of source code and related artifacts for conversion to run on newer hardware, operating systems and languages. The center of our proposed innovative approach is to map and parse a complex source code tree (plus related artifacts such as ‘makefiles’) into an appropriate process and object oriented schema tree. From this representation we propose to generate relational database queries, stored procedures and advanced algorithms that extract dependencies and partitioning with respect to hardware, operating system, distributed processing, application programming interfaces [APIs], etc. The proposed tools will help analysts visualize, analyze and gain a deep, multi-level understanding of legacy code. This approach supports automation of legacy code functional inventory, language translation and re-partitioning. The solution is extremely flexible and applies to multiple versions of Ada (ANSI/MIL-STD 1815A, ISO-8652:1987, ISO/IEC 8652:1995/Amd 1:2007). We also propose to develop a plug-in for a popular software tool such as Eclipse. This approach has potential to yield order of magnitude reductions in the man-hours/cost required to modernize Ada applications.