Contribution and Design Rules

The design science cycle of this thesis yields two kinds of transferable knowledge: a set of technical rules that prescribe how to act under stated conditions, and a smaller set of prescriptive principles that hold beyond the particular artefacts. Both are stated in the form a design science contribution requires. Each carries the situation it applies to, the action it prescribes, the reason the action is expected to work, and — the part that makes it a claim rather than an exhortation — the observation that would show it to be wrong. In the standard design-science vocabulary the contribution is primarily an improvement, a better way of representing housing adaptation than the established baseline, and secondarily an exaptation, in that representational-governance machinery developed for other domains is carried into housing adaptation. Both senses are scoped to representational governance for diachronic housing adaptation, and neither is claimed beyond it.

Technical Rules

The four technical rules are conditional prescriptions, and each is falsifiable: it closes with the observation that would refute it.

TR-01 — Make interface obligations explicit and serialisable. To preserve semantic intent across the handovers of a multi-actor housing-adaptation process in which change recurs, encode interface obligations and invariants as explicit, serialisable objects rather than leaving them tacit. Explicit obligations spare each later actor the tacit reconstruction they would otherwise perform, and they make every handover re-checkable. The rule holds across the representational pipeline of Chapters 5 to 9; it presupposes stable naming, versioned interfaces, and declared transformation units, and its cost is the higher upfront modelling effort and governance overhead that this discipline imposes. It is refuted if verification after a change still requires global reinterpretation despite the explicit obligations.

TR-02 — Bound verification by publishing trigger-to-check mappings. Where a modular architecture admits frequent small changes, publish the mappings from change-trigger to required check, tied to interface boundaries, so that the verification a local edit demands is finite and known in advance. A finite trigger set localises verification and keeps escalation debt from accumulating. The rule operates through the interface contracts of Chapter 6 and the change traces of Chapter 9; it assumes the trigger logic stays finite and the escalation policy is declared, and it risks missing hidden dependencies where boundaries are underspecified. It is refuted if local edits repeatedly trigger unbounded global checks.

TR-03 — Define a deterministic transformation grammar with replay diagnostics. To make transformation reproducible and auditable within a formal notation-and-parser workflow, define the transformation grammar deterministically and instrument it with replay diagnostics. Deterministic replay is what allows the validity of a transformation to be evaluated rather than merely asserted. The rule governs the formal core of Chapter 7 and the replay outputs of Chapter 9; it requires the parser semantics to be locked and the baseline comparator fixed, and it trades away flexibility for ad hoc, non-governed edits. It is refuted if equivalent inputs cannot be replayed consistently.

TR-04 — Scale variation through a shared rule library and a typed exception budget. To let a procedural generator vary its output under heterogeneous constraints without collapsing governance, enforce a shared rule library and grant variation only through a typed, budgeted set of exceptions. Entries of the governed instance library, of this kind, preserve compatibility while still admitting variation. The rule applies to the generation pipeline of Chapter 8 and the discussion of Chapter 9; it requires exception classes, budgets, and justifications to be pre-registered, and throughput can fall when the rule checks are strict. It is refuted if the variation gains depend on silent rule violations or untyped exceptions.

Prescriptive Principles

Three principles generalise beyond the specific artefacts to any comparably adaptation-heavy delivery chain.

DP-R1 — Separate semantic obligations from geometric rendering at the interface. Keeping what a representation means distinct from how it is drawn, at the point where artefacts meet, is reusable across any adaptation-heavy socio-technical delivery chain; it adds little where lifecycle change is negligible.

DP-R2 — Treat change as named transformation units with explicit invariants. Naming each change as a unit that carries its own invariants supports post-change assurance in any governed workflow, provided the setting maintains stable identifiers and version discipline.

DP-R3 — Use pre-registered comparators for burden and validity claims. Fixing the comparator before measuring is what gives a comparative design-science claim its force; the principle applies wherever utility is argued comparatively, and comparator drift invalidates the strength of any claim built on it.