Computable Design Intent: A Versioned Graph for Engineering Consistency and Evidence Traceability

Engineering decisions acquire meaning through relationships between requirements, rationale, configurations, authority, and evidence. These connections may be scattered across company records or remain implicit. Greater language-model capability can assist interpretation, but current organizational knowledge must still be supplied and maintained. Motivated by the author’s development of DVIT, this paper proposes evidence-backed ingestion into a persistent, domain-specific relationship map. Computable Design Intent (CDI) is the portion of intent represented with sufficient scope and semantics for a defined operation. The Computable Design Intent Graph (CDIG) connects these records with their sources, decisions, and history. The method makes initial capture and continuing maintenance central to downstream reuse. It can support retrieval, change review, constraint comparison, and evidence assessment, including through MCP. Semiconductor examples illustrate how scoped declarations and authored relationships yield review questions. The hypothesis to evaluate is whether reuse of this maintained context improves downstream work enough to justify its lifecycle cost.