State-Centric Framework for Canonical Representation and Change Analysis of Evolving Documents in Information Systems

Information systems increasingly operate on document collections that evolve continuously over time, are queried repeatedly, and must support traceable propagation of relevant changes. Despite this dynamic nature, documents are still commonly treated as static artifacts, which limits the explicit modeling of document evolution and may constrain meaningful comparison. This paper proposes a state-centric framework that reinterprets documents as structured and versioned states within an information space. Each document instance is represented as a canonical, hierarchically decomposed state linked to its contextually valid predecessor. On this basis, change is defined as a formally specified transition between comparable states rather than as an implicit difference between external artifacts. The framework integrates three interdependent layers: deterministic representation through canonicalization and hierarchical decomposition, temporal organization through explicit predecessor relations, and comparison mechanisms operating on structured states. A hierarchical digest-based realization is introduced as one computational instantiation of the framework. Its operational realizability is examined through a controlled and reproducible case study based on successive JSON document states, predefined correspondence information, and an executable reference implementation. Within this setting, equivalent document representations produced identical canonical digests, predefined structural changes were correctly detected and localized where valid correspondence information was available, and ambiguous component identities illustrated the limits of localization without persistent identifiers. Together, the analytical assessment and the reproducible operational case study provide complementary evidence for the conceptual coherence and operational realizability of the proposed framework under the stated representation assumptions.