The US Government spends upwards of $50 Billion a year maintaining legacy systems. Replacing these systems can be a challenge. Legacy software systems can be huge, evolving over time, developed and maintained by a different team every few years. There may be documentation, or there may not be. If there is, it may be poor quality or unmaintained. Cyber security and data breaches are difficult to address in these systems. And system interoperability is especially challenging in domains where mergers and acquisitions create overlaps in functionality that must be deconflicted. NLP is a rapidly advancing field which describes a wide umbrella of technologies and techniques useful in traversing, indexing, classifying, and analyzing large amounts of unstructured textual data in order to aid and speed inquiries about that data. However, the tools for analyzing software source code have been stagnant for many years. Praeses builds upon recent research in academic and commercial spheres to prove the potential of applying NLP techniques against software languages. By using these techniques to process software and generate metadata, the resultant data can be leveraged by existing tools and form the foundations for new toolsets to dramatically improve outcomes in systems modernization, reverse engineering, and cyber security.
