A technique to recover an unknown rational number p/q from its modular image r modulo an integer M, typically by finding small numerator and denominator consistent with r via continued fractions or lattice-based methods.

Principle

Combine modular residues with bounds on numerator and denominator and either continued-fraction approximants or lattice reduction to identify a unique rational whose modular image matches the given residue under size constraints.

Demonstration

Given r ≡ p·q^{-1} (mod M) and bounds |p|≤P, 0

Misapplication

Applying rational reconstruction without verifying size bounds or modulus magnitude: attempting to reconstruct when PQ is not sufficiently smaller than M can produce false rationals or multiple candidates, misleading downstream algebraic computations.

Consequence

When the prescribed bounds and modulus conditions hold, rational reconstruction produces exact rational coefficients from modular computations, enabling correct lifting of modular polynomial factorizations and integer-resultant computations to Q or Z.

Reversal

The inverse is treating arbitrary residues as unrecoverable reals: accepting modular residues as opaque and refusing to attempt reconstruction loses the opportunity to recover exact rational data from modular calculations.

Boundary

Valid when there are proven or credible bounds on numerator/denominator and modulus M large relative to PQ; excludes ambiguous cases with multiple short candidates or noisy residues from arithmetic errors or noninvertible denominators modulo M.

Semantic Tension

Tension exists between continued-fraction approaches and lattice-based reconstructions; both aim to recover rationals but differ in robustness to noise, size limitations, and ease of proving uniqueness.

Synthesis

Rational Reconstruction is the process of combining a modular residue with size bounds and either continued-fraction approximants or lattice reduction to recover a unique rational number p/q whenever modulus and bound conditions guarantee uniqueness.