Treating power-of-two scaling as a complete numerical repair

Exact statement

Treating power-of-two scaling as a complete numerical repair. it removes coefficient representation error in an unbounded-exponent binary model but does not prevent amplification of earlier roundoff by later $|t|^{-3}$ scaling, nor does it address finite exponent range, overflow, or underflow.

Statusrejected.

External reviewNo documented external or specialist review of the FPRD synthesis is recorded.

Context

The research record preserves failed approaches when they expose a model boundary, an invalid transfer, or an unsupported strengthening.

Proof or evidence

it removes coefficient representation error in an unbounded-exponent binary model but does not prevent amplification of earlier roundoff by later $|t|^{-3}$ scaling, nor does it address finite exponent range, overflow, or underflow.

Rejected transfers and surviving boundaries →

Verification notes

The failure status and preserved lesson are recorded in the cumulative failure ledger.

Limitations

representation, roundoff propagation, and exponent range are separate numerical receipt fields.

Open work

Keep this exclusion attached to any future reuse of the corresponding construction.