A new arXiv paper introduces PULSE, an executable contract language for spatiotemporal knowledge graph engineering. The system is meant to keep accepted state, observations, constraints, processes, and hypothetical scenarios inside one typed runtime rather than scattering their execution rules across artifacts.
PULSE defines operational roles and their write effects, including evidence non-overwrite, branch isolation, guarded state changes, timers tied to grounded subjects, and ordered events over time and space. Generated views remain compatible with standards such as GeoSPARQL, SOSA, and SHACL.
The authors also describe formal and implementation checks, including Lean 4 kernel analogues, bounded tests, and runtime-kernel cases. Those details matter because knowledge graphs used for operations need predictable behavior, not only expressive schemas.
The work is highly technical, but its practical aim is clear. As knowledge graphs model changing real-world systems, teams need enforceable contracts for what can update, when, and on what evidence.