7ccede73474108f17d774ff6edc99c6f07b20711. Identity is necessary for a receipt, but it is not proof of the page.
Jill · exploratory model · evidence boundary
Six-axis gemstone study
The earlier “perfect stone” story combined subjective scoring, optical and material constants, an optimizer, synthesis ideas, and source excerpts into a single verified-sounding conclusion. The evidence does not support that conclusion, so the interactive claim surface is paused.
- Six-axis optimizerSubjective weights, empirical constants, and browser sweeps
- Archived“Perfect stone” and global optimum claims removed
- Reopen with receiptsMeasured inputs, reproducible runtime, and exact theorem scope
No stone is declared perfect or optimal
A Pareto visualization depends on its chosen axes, normalizations, weights, candidates, and numerical implementation. A source theorem about a typed score or a conditional model identity does not validate empirical inputs, manufacture a material, or prove global physical superiority.
Current evidence ledger
sorry markers and declared axioms are parser results, not a compiler or kernel receipt.
BeautyWeights and BeautyAxes types, weightedSumBeauty_le_one proves an upper bound. It does not turn a subjective score into a probability or select a real material.
Claims removed from the product surface
- Whole-page statements that every number, tradeoff, or optimizer parameter was machine-checked.
- The inference that a conditional factorization or a hard-coded numeric inequality validates physical independence, correlation, dominance, or Pareto optimality.
- Empirical claims about rutile, diamond, photonic crystals, fluorescence, durability, synthesis, SEM observations, and finished-stone appearance without a versioned evidence packet.
- Stale theorem/declaration counts and broad “zero unproved claims” language derived from source parsing.
- Manufacturing guidance implied by exploratory CVD, coating, carrier, or laboratory descriptions.
Path back to an interactive study
Reopening requires sourced material measurements with units and uncertainty, explicit subjective-value choices, a validated optical/material model, a reproducible optimizer packet tied to an exact implementation commit, exact theorem bindings with their full hypotheses, and UI tests that prevent structural inventory from appearing as kernel or empirical certification.