This paper presents an open-source Tree-sitter grammar for MUMPS, the programming language used by VistA, the U.S. Department of Veterans Affairs' electronic health record system. Existing MUMPS parsers are typically embedded within language runtimes and designed for execution rather than analysis, preventing the use of modern developer tooling on large MUMPS codebases. The grammar generates a concrete syntax tree suitable for static analysis and tooling. It successfully parses 98.6% of VistA's 33,951 routines (approximately 3.6 million lines of code) without producing error nodes. A differential testing framework comparing parser behavior against YottaDB validates the correctness of expression parsing and semantics. Using this parser, the paper performs the first large-scale structural analysis of VistA. It reconstructs the system's call graph, showing that nearly all packages depend on the foundational FileMan and Kernel subsystems, and quantifies the prevalence of runtime code generation, providing empirical insight into the practical limits of static analysis for the VistA codebase.
A Srivastava (Sun,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: