← Instrument galleryD06-SURFACEDELTA-VERSUS-PROSE

SurfaceDelta Versus Prose

A deterministic browser-local research instrument.

DETERMINISTIC FIXTURE · NOT A PERFORMANCE RESULT

SurfaceDelta Versus Prose

MECHANISMSchema-bound graph patches over an exact source surface
AUTHORITYVersioned Semantic Surface Graph
COMMIT RULEBase version, targets, budgets, and citations validate before render.
FAIL-CLOSED RULEStale or unsupported deltas reject without partial mutation.
STATEVALID23471fb0
base version0
repair countVALID
citation state23471fb0
  1. VALID
  2. INVALID_JSON
  3. STALE_BASE
  4. CITATION_BREACH
  5. REPAIRED
  6. VERIFIED_OR_FLAGGED
CONTRACT CONTROLS

Each control executes a declared browser-local transition. Repeated dual-outcome controls alternate deterministically.

FIXTURE ASSERTIONS

NOT YET EXERCISEDEvery delta validates before render.

NOT YET EXERCISEDStale base rejects without partial mutation.

NOT YET EXERCISEDEvery quote is verified or visibly flagged.

NOT YET EXERCISEDMetrics separate latency, output tokens, repairs, and usefulness.

RECEIPTED EVENT LOG

No transition yet. Operate a declared control.

This transparent browser-local instrument teaches the authored state contract. Exercising an assertion demonstrates fixture behavior only; it is not evidence of production performance, compression ratio, model uplift, cognition, safety, or universal correctness.

Purpose

Compare full-answer generation with bounded graph patches under identical source, task, and model fixtures.

Mechanism

Condition A asks for a complete brief. Condition B asks for typed SurfaceDelta operations against an exact base version. The deterministic renderer and quote verifier evaluate both paths.

Deterministic fixture

A source-backed contract risk analysis with exact spans, five surface blocks, one stale base version, and one intentionally malformed citation.

The fixture must live in content/demos/d06-surfacedelta-versus-prose/fixture.json, validate before the demo renders, and be resettable without a network call. A future live adapter may be added behind a visibly separate mode.

Required controls

  • Choose condition
  • Run deterministic response fixture
  • Apply patch
  • Inject stale version
  • Break citation
  • Repair once
  • Show rendered result
  • Compare metrics

Visible states

  • VALID
  • INVALID_JSON
  • STALE_BASE
  • CITATION_BREACH
  • REPAIRED
  • VERIFIED_OR_FLAGGED

Every state must have text, iconography, and a non-color-only distinction. State transitions are logged in an in-memory demo receipt visible in the inspector.

Core assertions

  1. Every delta validates before render.
  2. Stale base rejects without partial mutation.
  3. Every quote is verified or visibly flagged.
  4. Metrics separate latency, output tokens, repairs, and usefulness.

Layout

The page contains:

  1. a concise mechanism explanation;
  2. the interactive stage;
  3. a state and evidence inspector;
  4. a reset control;
  5. a “what this proves / what it does not prove” panel;
  6. links to the related paper and benchmark contract.

On narrow screens the inspector becomes a bottom sheet. The stage must remain usable at 360 CSS pixels.

Accessibility

  • All operations are keyboard reachable.
  • Pointer gestures have button or keyboard equivalents.
  • Animated transitions honor prefers-reduced-motion.
  • Focus remains visible and returns predictably after sheets close.
  • Diagrams expose concise text alternatives.
  • Status is announced through a polite live region only when the user initiates the transition.

Instrumentation

Record locally:

demo_id
fixture_version
action
prior_state
result_state
validation_result
timestamp_relative

Do not send telemetry. Provide “Export demo trace” as JSON for debugging and reproducibility.

Automated tests

  • fixture schema validation;
  • deterministic reset;
  • every required control transition;
  • every assertion above;
  • invalid and stale fixture handling;
  • keyboard path;
  • 360 px and 1440 px screenshots;
  • zero critical console errors;
  • no request to a paid API in default mode.

Publication boundary

A real model uplift claim requires matched local-model runs; fixture mode demonstrates the benchmark contract only.

The interface must never convert illustrative values into a benchmark claim. A measured result appears only when linked to a versioned run receipt.