VOLE-Document: Procedural Text and Document Storage by Deterministic Structural Materialization, Entropy-Coded Reconstruction State, and Residual-Exact Replay - Broad Prior-Art Technical Disclosure and Research Architecture

Abstract Digital text and document systems usually treat a serialized byte sequence as the durable primary object. Plain text is persisted as literal encoded symbols; JSON, XML, HTML, Markdown, source code, and logs are persisted as concrete syntax; office documents are persisted as package members, markup trees, relationships, styles, media, and ZIP metadata; PDF persists a graph of objects, streams, cross-references, page-description operations, embedded resources, and potentially multiple incremental revisions. In each case, a large fraction of the observed bytes may be the materialized consequence of simpler structure: grammars, templates, dictionaries, repeated subtrees, style systems, counters, tables, relationships, version deltas, compressor decisions, and serializer conventions. VOLE-Document is a proposed native, deterministic representation architecture that treats those observed bytes as a materializable view rather than assuming that they must always be the primary persisted form. Its exact profile represents a document as a bounded reconstruction description containing procedural structure, explicit parameters and state, entropy-coded event and residual streams, serializer/package reconstruction information, integrity metadata, and an unconditional raw fallback. The governing invariant is uncompromising: Materialize(descriptor) == original_bytes for every object admitted to the byte-exact profile. Semantic equivalence, canonical equivalence, visual equivalence, or successful parsing are not substitutes for this invariant. The central hypothesis is not that documents can be generated from a magical short seed, nor that information theory can be bypassed. The hypothesis is narrower and testable: many document byte streams contain structure that can be represented more cheaply as deterministic reconstruction procedure plus parameters and residuals than as a conventional generic compressed stream. rANS is used as an entropy-coding substrate for the remaining typed symbol streams; it is not itself the procedural model. The term rANS entropy seed is therefore defined operationally in this disclosure as a complete decoder-entry entropy capsule—model identity or model bytes, normalized frequency tables where required, rANS state, renormalization payload, symbol count, and integrity bindings—not as a claim that one integer state alone recreates an arbitrary document. VOLE-Document explicitly composes and extends ideas with substantial prior art: Lempel–Ziv parsing, statistical and context modeling, grammar compression, tree grammars, EXI-style grammar-driven structured encoding, delta compression, content-addressed storage, exact recompression of already-compressed streams, predictive coding, ANS/rANS, and minimum-description-length model selection. It does not claim invention of those components. The disclosed contribution is the broad systems architecture that places them under a single exact reconstruction contract, makes residuals first-class typed evidence, prices every representation by complete physical cost, permits residual-guided deterministic search without giving search heuristics decoder authority, supports standalone and store-backed forms, and requires falsifiable empirical courts before any mechanism is promoted. The architecture is intended to interoperate conceptually with the existing VOLE, EntropyFS, DSFB, and ryg-rans-rs stack. VOLE contributes the doctrine of storing deterministic explanation plus explicit residual and materializing the conventional observation on demand. EntropyFS contributes representation-aware persistence and exact logical-byte materialization. DSFB contributes a deterministic residual-observer role suitable for search-budget allocation and residual-structure discovery without entering the normative decode path. ryg-rans-rs contributes a native Rust rANS implementation and an evidence-oriented parity methodology. These relationships are architectural, not an assertion that VOLE-Document measurements already exist. Keywords document storage; procedural documents; exact reconstruction; rANS; residual coding; grammar compression; structured documents; PDF; DOCX; EntropyFS; DSFB; prior art