Requirements Traceability and Software Quality in Enterprise Application Development

Abstract

Purpose: Enterprise teams invest heavily in traceability, yet the claim that trace links improve software quality rests on thinner evidence than practice guidance assumes. This review treats trace links as an engineered asset with a creation cost, a recovery error, a decay rate and a measurable return. Design: A protocol-driven integrative review of requirements engineering, software maintenance, testing and assurance literature published up to 2024 addresses four questions: how link types, granularity and traceability information models shape utility; how accurately information retrieval and learning methods recover links; how links decay and what upkeep costs; and what evidence connects traceability to defect density and change impact accuracy. Findings: Vector space and latent semantic indexing recovery reaches useful recall only at precision levels that move cost into analyst vetting, and benchmark results transfer poorly to enterprise repositories. Link value concentrates in volatile, high-criticality requirements, whereas uniform coverage mandates produce stale links whose false assurance can be worse than an honest gap. Evidence relating completeness to lower defect density is directionally consistent but largely correlational. Contribution: The Trace Asset Value Model defines link quality through completeness, correctness, currency and granularity fit, routes it through impact analysis, test selection and audit channels, and states six falsifiable propositions with tests. Implications: Maintenance effort should follow change exposure and criticality, and link currency deserves the audit attention that coverage receives today.