Cross-Volume Reference Architecture for the Document Engineering Series: A Validation Framework

The Document Engineering textbook series for Vietnamese learners aged 14 and up (Vol 0–5: Foundation, File Systems, Markdown, LATEX, Citations, Diagrams) proposes a cross-volume reference matrix in ARCHITECTURE.md §7. The matrix is forward-design: it declares which volume sections reference which other sections across the series. After Vol 0– 2 have been drafted and Vol 3–5 are scheduled, the matrix needs post-hoc validation against the realised text. This paper specifies the validation framework. The framework distinguishes four cross-reference types — intra-volume, cross-volume, externalresource, and self-reference loops — and defines for each type a target-existence rule, a semantic-fit rule, a density rule guarding over- and under-linking, and a drift-detection rule that triggers re-validation when source or target prose changes. Seven named validation checks operationalise the rules at chapter scale: Target-Exists, Semantic-Fit, Density-Within-Band, No-Forward-Reference-Cycle, Single-Source-of-Truth, VolumeBoundary-Respected, and Refresh-Trigger-Defined. Each check is presented with definition, evidence required, characteristic failure mode, and fix recipe sized for a single editorial pass. The validation workflow is post-volume-drafted, batch-mode, auditloop-integrated; a sample completed validation summary across a hypothetical Vol 0–2 state demonstrates the report format. The framework is grounded in information architecture (MorvilleRosenfeld), academic cross-reference discipline (Plaut 1996), book-design canon (Williams, Bringhurst), technical-writing style guides (Microsoft Manual of Style, Chicago Manual of Style), wiki-link integrity literature (Hennessey-Ge 2013, Halfaker et al. 2013), and citation reference-rot research (Klein et al. 2014, Spinellis 2003, Zhou et al. 2015). Limits, manualversus-automated trade-offs, and four open questions on AIassisted validation are discussed. The framework's primary downstream effect is that every volume's textual cross-references — internal callbacks, forward pointers, cross-volume connectives, and external resource links — gain a defined pass-fail criterion that the audit loop can score.