Two Kinds of Guarantee: Hybrid Primitives, Sample-Oracle Restriction and One Misplaced Attack in Mamun's "Theoretical and Physical Boundaries of Lattice-Based Post-Quantum Cryptography" (Vol-II)
Year: 2026
Type: article
Abstract: This review examines Syed Muntasir Mamun's "Theoretical and Physical Boundaries of Lattice-Based Post-Quantum Cryptography: An Anatomical Critique of Hybrid Info-Theoretic and Computational Primitives (Vol-II)" (2026). The paper organises a large and scattered literature around a distinction that genuinely matters and is often blurred: computational security and information-theoretic security are not two strengths of the same guarantee but two different kinds of guarantee, and a system that needs permanence cannot obtain it from a hardness assumption alone. On that axis the paper assembles a coherent taxonomy of hybrid primitives — the Lattice PUF binding modelling resistance to lattice hardness through learning theory, quantum-locked physical primitives deriving authentication from conjugate coding rather than from algorithmic difficulty, key-combining protocols in the Muckle family, and leakage-resilient treatment under the Bounded Retrieval Model — and it keeps the discussion connected to deployment constraints rather than to idealised channels. Its treatment of quantum sample oracles is the analytical high point and rests on a correctly identified and correctly characterised result. One substantial correction is recorded. The paper's framing claim, which appears in the abstract, is that lattice cryptography is vulnerable to Gottesman-Kitaev-Preskill error correction. GKP is a bosonic quantum error-correcting code: it encodes a logical qubit in the phase space of a harmonic oscillator and corrects small displacements of that oscillator, performed by a party who physically holds the quantum state. It is not a cryptanalytic algorithm against classical ciphertext, no polynomial-time quantum algorithm for the lattice learning problem is known, and the defence proposed against the supposed attack would, read literally, also prevent legitimate decryption. The thesis does not depend on that claim and is better supported without it. Further findings concern the bibliography, which repeats one source in three link forms, carries two entries with no visible relation to the subject, and gives no authors, venues or persistent identifiers. The taxonomy is the contribution, and it survives every finding recorded here.
Author: Islam, Md. Saiful
Language: English
Subfield: Computer Networks and Communications
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7236161563
