Glossary and notation

Core terms

Core Technology vocabulary. Short definitions here; formal depth lives on the linked pages.

Channel ZeroSecurity

An authorization transcript where the public record reveals nothing about the hidden secret, given the public challenge. Capture is not knowledge: an eavesdropper can archive the channel and still learn zero about the private state under the stated model.

Formal note: Model claim: $$I(X; O \mid C) = 0$$: mutual information between hidden state $X$ and observable transcript $O$, conditioned on public challenge $C$, is zero under the named axioms.

See also: When the channel reveals nothing · Math hub

mutual informationMath

A measure of how much one random variable reveals about another. Perfect secrecy and Channel Zero are zero-mutual-information claims: observing the public transcript does not shrink uncertainty about the hidden secret.

Formal note: Defined as $I(A;B) = H(A) - H(A\mid B)$ (and the conditional form $I(A;B\mid C)$). Channel Zero targets $I(X;O\mid C)=0$; Shannon perfect secrecy for messages is the related claim $I(M;C)=0$.

See also: Notation: I(·;·) · Perfect secrecy

perfect secrecyMath

Shannon's gold standard: after seeing the ciphertext (or public transcript), your belief about the secret is unchanged. Security that does not rely on computers being too slow, only on the mathematics of the channel under stated assumptions.

Formal note: Equivalences (message setting): $I(M;C)=0$, or $H(M\mid C)=H(M)$, or posterior equals prior for every ciphertext with positive probability. OTP with uniform one-time key is the classic existence proof.

See also: Perfect secrecy (Shannon) · Computational bargain

unicity distanceMath

How much ciphertext (or archive) is needed before language redundancy collapses an ordinary cipher to one plausible reading. Explains why “more data” often helps attackers against computational schemes, and why Channel Zero aims to deny that collapse on the public channel.

Formal note: Shannon’s classic estimate $u \approx H(K)/D$: key entropy over per-symbol redundancy: marks when expected spurious keys fall to one. Authorization “anti-unicity” is the opposing design goal for transcripts (see anti-unicity).

See also: Why more data doesn’t help · Notation: u

surface bitsMath

Bits an observer can actually measure from the public channel or ledger record. Surface bits are what archives hold; they are not the same as the secret’s effective information.

Formal note: Often written $H_{\mathrm{surface}}$: entropy of the observable alphabet or archive. Nonzero surface entropy is compatible with $I_{\mathrm{effective}}=0$ when marginals are flat and independent of the secret under the model.

See also: What observers can measure

effective bitsMath

The useful secret information that still matters for authorization after accounting for what the public transcript exposes. Observer-bit arguments compare surface bits to effective bits to show what an archive can never learn.

Formal note: Effective leakage is mutual information with the secret (e.g. $I(X;O\mid C)$) not raw archive size. Observer Bit Model claims keep $I_{\mathrm{effective}}=0$ while $H_{\mathrm{surface}}$ may be positive.

See also: What observers can measure

bearingProtocol

A public label emitted instead of revealing the underlying private choice. Bearings are one-shot synonyms on the channel: by the time anything could observe them, the ceremony has already moved on.

Formal note: Public witness symbols $W$ (or $\upsilon$) drawn from a finite bearing alphabet $B$. Uniformity axioms ask that, given $C$, each label is equiprobable: $P(W=w\mid C)=1/|B|$.

See also: How Rosario works · Notation: bearings

private map φMath

The committed private mapping from hidden zones to the public bearing labels that appear on the channel. Losing φ is an endpoint problem; Channel Zero claims are about what the public transcript reveals, not about a stolen laptop.

Formal note: Denoted $\varphi$: a private bijection (or committed morphism) from hidden zones / synonyms to public bearing labels. Channel claims concern $I(X;O\mid C)$; compromise of $\varphi$ is an endpoint threat outside that claim.

See also: Notation: φ · Channel Zero

twin circuitsIdentity

A pair of related circuit identities used so proof and verification can agree without exporting a reusable authority object on the channel. Twins keep identity capability-shaped rather than copyable-token-shaped.

Formal note: Matched prover/verifier pair sharing a sealed private relationship (compiled twins). Membership outcomes agree locally; the private map never appears as a reusable bearer artifact on $O$.

See also: Identity in twin circuits

burn-before-validateLedger

Spend the one-shot nonce on the ledger before cryptographic validation runs. Replay races die at the systems layer: every submission consumes the nonce even if the proof later fails.

Formal note: Operational latch for session-bound no-replay (axiom P4): nonce state transitions RESERVED→SPENT before proof validation. Replay races fail at the ledger, not only at cryptographic rejection.

See also: Recording without learning · How Rosario works

computational bargainSecurity

Today’s mainstream crypto bet: attackers are assumed too slow to break hard math problems. It works until algorithms, hardware, or archives catch up. Information-theoretic authorization does not make that bet for the channel transcript.

Formal note: Security reductions to problems believed hard (factoring, discrete log, lattices, etc.). Distinct from information-theoretic emptiness $I(X;O\mid C)=0$, which does not rely on a hardness clock for the transcript claim.

See also: Today's crypto: a bet on hard math · Compare security systems

structured ambiguitySecurity

Deliberate design so many secrets remain compatible with the same public transcript. Ambiguity is not a bug: it is how the channel stays unreadable even when observers collect more data.

Formal note: Preserves observer equivocation: many hypotheses remain consistent with $O_{1:n}$. Related to anti-unicity: archive growth need not drive residual key uncertainty to zero under the model.

See also: Ambiguity by design

observerSecurity

Anyone who can see public challenges, bearing witnesses, or ledger records, without the private map. Observer arguments ask what that viewpoint can measure, not what an endpoint thief can steal.

Formal note: Viewpoint that receives $(C,O)$ (and ledger surface) but not $\varphi$ or local twin state. Formal claims quantify $I(X;O\mid C)$ for this viewpoint, not endpoint theft.

See also: Recording without learning · Observer bit budget

challengeProtocol

The public context a prover must answer in a ceremony round. In Channel Zero notation it is often written C; it is visible, but under zero mutual information it does not leak the hidden state X.

Formal note: Public context $C$ in Channel Zero notation (disambiguate from classical ciphertext $C$ in encryption prose). Conditioning in $I(X;O\mid C)$ treats $C$ as known to the observer.

See also: Notation: C · Channel Zero

envelopeProtocol

The authority object that replaces a bearer token. A one-shot proof bound to the method, the endpoint, the hash of the request body, the policy, the timestamp, and a single-use nonce, so it authorizes exactly one request and nothing else.

See also: Engineering stack

reference architecturePlatform

The public mechanism map for how gates, ledgers, and circuits fit together. Absolute mechanism claims on the site point here so buyers can see the intended shape without deployment threat detail.

See also: Reference architecture

information-theoretic authorizationSecurity

Authorization whose channel guarantees do not rely on computers being too slow. Security is argued from entropy and mutual information on the public transcript: Shannon’s paradigm applied to access, not only to message secrecy.

Formal note: Applies Shannon-style emptiness to access transcripts: channel guarantees argued from $H(\cdot)$ and $I(\cdot;\cdot)$ on $(C,O)$, not from a computational hardness clock for that claim class.

See also: Why not all proofs are equal · Claims

Living glossary

Terms used across Fundamentals, Proofs, Math, and Reference. Hover tooltips and deep links land on anchors on this page.

anti-unicityMath

The design goal opposite Shannon unicity for authorization transcripts: more observations do not collapse the hypothesis space to one secret. Balanced zones, private maps, and fresh rounds preserve observer-side non-unicity under the model.

Formal note: Opposite of Shannon unicity collapse for authorization transcripts: $O_{1:n}$ need not uniquely determine $X$ as $n$ grows. Achieved via balanced partitions, private $\varphi$, and session independence under the axioms.

See also: Unicity distance · Ambiguity by design

artifact problemSecurity

When a downloadable proof object (a token, signature blob, or zero-knowledge proof file) can be copied and reused as ambient authority. The artifact travels; possession of the file becomes capability again. Channel Zero ceremonies aim to avoid exporting that kind of reusable proof.

See also: Why not all proofs are equal

attack conversionSecurity

The step where an attacker turns something visible or stealable (transcript, film, vault dump) into usable secret knowledge or reusable authority. Classical families fail when that conversion is denied by empty-channel design and burn-before-validate.

See also: Why classical attacks fail

bearingsProtocol

The six directional choices (up, down, left, right, forward, back) a prover makes each round. Bearings are one-shot: by the time anything could observe them, they are already spent.

See also: How Rosario works · Notation: bearings

binaryArtifact

The compiled executable form of a circuit. Built from the standard, hardened, or hybrid source with its configuration and entropy embedded, so identity travels with the artifact rather than sitting in a vault.

burnLedger

Spending a nonce so it can never be used again. The gate burns the nonce before it validates the proof, which is the structural reason replay fails: every submission spends the nonce even when validation later rejects it.

See also: Recording without learning · Security

capture ≠ compromiseSecurity

Recording a Channel Zero ceremony (or stealing the public transcript) does not grant usable authority over the secret under the model. Unlike a stolen password or bearer token, the archive is not ambient capability.

Formal note: Capture of $O$ is not possession of ambient authority when $I(X;O\mid C)=0$ and one-shot burn semantics hold: distinct from stolen passwords or bearer tokens.

See also: Channel Zero · Claims

challenge geometryProtocol

The fresh public layout of zones or symbols presented each round. The prover does membership work against this layout; the private map never leaves the local twin. New entropy each session keeps yesterday’s answers from authorizing tomorrow.

See also: How Rosario works · Axioms

circuitIdentity

A user's or workload's cryptographic identity, embodied as a compiled binary that carries its own config, entropy, and policy. Identified by a handle and by the SHA-256 of the artifact.

See also: Twin circuits

cognitive proofSecurity

A human-facing proof of knowledge through perceptual membership work (recognize and choose under a fresh layout) without typing a reusable secret onto the channel. Presence is shown as work, not as a biometric template dump.

See also: Human ceremony proofs

commitLedger

The ledger operation that marks a nonce SPENT.

compiled twinProtocol

One half of a matched prover/verifier pair that shares a sealed private relationship. Identity deploys as this pair, not as a password row. The twins agree on membership outcomes without putting the private map on the channel.

See also: Twin circuits · Engineering stack

computational sealingSecurity

Using hashes, AEAD, and XOF streams to protect integrity and confidentiality of artifacts. Operational crypto, not the same as an information-theoretic claim that the authorization channel carries zero secret-relevant information.

See also: Engineering stack · Claims

entropyMath

Uncertainty measured in bits. High entropy means many equally plausible outcomes; low entropy means the answer is almost decided. Conditional entropy is what remains after an observation: perfect secrecy keeps that remainder equal to the prior.

Formal note: Shannon entropy $H(A)=-\sum_a P(a)\log_2 P(a)$; conditional $H(A\mid B)$. Perfect secrecy keeps $H(M\mid C)=H(M)$.

See also: Notation: H(·) · Math hub

entropy poolArtifact

A sealed store of freshness material bound to a circuit binary. Starved, reused, or tampered entropy voids ceremony assumptions; integrity failure detection is an operational property of the engineering stack.

See also: Engineering stack · Claims

EPINPlatform

The public API surface: the set of public domains and endpoints through which the platform is reached.

equivocationMath

Leftover uncertainty about which secret produced the public transcript. In Channel Zero designs, structured ambiguity keeps many hypotheses alive for the observer (even after a perfect recording) so uniqueness never arrives.

Formal note: Residual uncertainty about the secret after observing the transcript, related to conditional entropy $H(X\mid O,C)$. Anti-unicity designs aim to keep observer equivocation from collapsing as archive length grows.

See also: Ambiguity by design

four planesPlatform

Foundry (mint), Control (registry and ledger), Gate (enforce), and Verify (adversary harness): the product architecture that ships empty-channel authorization without restating Shannon on every product page.

See also: Engineering stack · Products

gate orderProtocol

The fixed sequence of checks a gate runs on an untrusted request before any application logic executes: recompute the request hash, check endpoint and policy, confirm the circuit is active, check freshness, burn the nonce, validate the proof, apply application policy, then serve.

See also: Security

Grover's algorithmSecurity

A quantum unstructured-search algorithm offering roughly a quadratic speedup over classical guessing. Powerful against large hypothesis spaces, but not an exponential free lunch, and useless at inventing a secret that was never in the observation.

See also: Quantum myths vs real bounds

handleIdentity

The user or organization identifier that resolves to the current circuit. Deactivating a handle is how revocation happens (one state change rather than a rotation campaign).

hardness clockSecurity

A metaphor for computational security: protection lasts only while hard math problems stay too expensive to break. Machines, algorithms, or archives can run the clock down. Information-theoretic channel claims do not make that bet for the public transcript.

Formal note: Label for security lifetime tied to computational assumptions (and harvest-now-decrypt-later archives). Contrasts with information-theoretic transcript claims that do not expire when a hardness problem falls.

See also: Computational bargain · Math hub

harvest now, decrypt laterSecurity

An attacker records encrypted traffic today and waits for cheaper compute, better algorithms, or a quantum break tomorrow. Sessions end; the archive keeps ticking. Perfect forward secrecy shortens the window: it does not make the channel empty of secret-relevant information.

See also: Computational bargain

hidden state XMath

The private secret state a Channel Zero ceremony is designed not to leak, often the private map and related entropy. Notation writes X for what stays off-channel; O for what observers see; C for the public challenge.

See also: Notation: X · Channel Zero

honesty wallSecurity

The public scope page that separates what ENI6MA claims from what it does not. Absolute language on Technology pages should point here so buyers can see assumptions and failure modes.

See also: Claims

hybrid circuitIdentity

A circuit variant with a four-shard prime and an X25519 + Kyber768 key exchange, giving post-quantum protection for the key agreement. Adds agent mode for machine-to-machine use.

job fitSecurity

Choosing a security paradigm for the threat it actually addresses (transport secrecy versus authorization emptiness) rather than declaring one system “best crypto.” Matrices teach fit; they do not crown a winner for every cell.

See also: Compare security systems

language redundancyMath

Structure in ordinary language (or any non-uniform source) that lets analysts discard impossible keys as ciphertext grows. Shannon’s unicity distance shrinks when redundancy is high: why “more data” often helps against classical ciphers.

See also: Unicity distance

manifestArtifact

The metadata record for a binary: its hash, size, timestamps, and self-validation state. The registry uses it to verify and index artifacts.

membership witnessProtocol

Public evidence that the prover did membership work on a fresh geometry or challenge, without exporting the private map itself. Presence is shown by doing the work now, not by presenting a stored password string.

Formal note: Public evidence $W$ that a zone/leaf membership check succeeded for the fresh challenge geometry, without exporting $\varphi$. Soundness is combinatorial over ring size, rounds, and balanced partitions; the witness is not a reusable credential string.

See also: How Rosario works

model-scopedSecurity

A strong claim that holds only under named axioms, observer models, and deployment patterns (such as burn-before-validate), not as ambient marketing physics that ignore side channels or skipped architecture.

Formal note: Absolute words on Technology pages inherit scope from named axioms (e.g. P1–P4), observer models, and deployment latches, not ambient physics. See Claims for the affirmative model-scoped table.

See also: Claims · Axioms

nonce_uuidLedger

The identifier of a ledger-reserved nonce. It is issued at reserve, returned with the proof, and spent on use. A nonce is good for exactly one request.

Observer Bit ModelMath

Rosario–Wang framing that separates surface bits (what an archive can record) from effective bits (mutual information with the secret). Capture fills notebooks; under the axioms the effective quantity stays zero.

See also: Observer bit budget

observer's ledgerSecurity

A perfect archive of everything visible on the public channel: challenges, bearings, timestamps, verdicts. The notebook can fill with surface bits while effective knowledge of the secret stays empty under the stated observer model. Capture is not compromise.

Formal note: Idealized archive of $(C_i, O_i)$ (and ledger surface) across rounds. Formal claim: archive growth need not raise $I(X;O_{1:n}\mid C_{1:n})$ above zero under the axioms.

See also: Recording without learning

one-shot proofSecurity

A proof of knowledge bound to a single request (or freshness epoch) so the authority it carries dies with that moment. Unlike a reusable password, token, or certificate, a one-shot proof is not ambient capability an attacker can steal and replay later.

See also: Technology overview · How Rosario works

one-time padMath

The classic perfect-secrecy cipher: mix the message with a truly random key of equal length, use it once, and never reuse it. Existence proof that emptiness is possible: logistics of distributing pads at internet scale are why most systems chose computational ciphers instead.

Formal note: Vernam/OTP: $C = M \oplus K$ with $K$ uniform, independent, equal length, and one-time. Yields $I(M;C)=0$. Existence proof for perfect secrecy, not a claim that OTP logistics scale to internet authorization.

See also: Perfect secrecy · Notation: ⊕

parameterized boundsMath

Public reference numbers for the empty-channel model: surface entropy per bearing, effective mutual information at zero, ring sizes, and bearing cardinality |B|. Labels for product claims, not fantasy forever-guarantees.

See also: Axioms

path-complete geometric projectionMath

Axiom P1: the public challenge enumerates every selector path any secret in the family could take. Completeness keeps the choice of walk from leaking by itself, partial menus would bias an observer.

See also: Axioms

per-event workSecurity

Membership labor done for one freshness epoch, fresh challenge, fresh witnesses, spent nonce. Authority dies with the event. Cloning a bearer token is easy; forging matching twin work without the sealed map is a different game.

See also: Twin circuits · Human ceremony proofs

perfect forward secrecySecurity

A transport property: long-term keys cannot decrypt past sessions if ephemeral handshake secrets are discarded. Valuable mitigation for TLS. It is not Shannon emptiness: stolen bearer tokens inside a tunnel still authorize until revoked.

See also: Computational bargain

presence as workSecurity

Showing you are here by doing membership decisions now, not by presenting a stored face, fingerprint, or password string. Deepfake-resilient framing aims at the work, not at archiving biometric templates.

See also: Human ceremony proofs

proof_hashProtocol

The hash the circuit produces from a proof run, carried in the proof payload and checked during validation.

protected endpointPlatform

A single route or tool placed behind a gate. It is the unit of value on every commercial page (what you buy is endpoints under proof, not seats or circuits).

public papers corpusPlatform

The site-indexed set of public ENI6MA white papers and essays with full abstracts and audience tags. Primary sources for Technology, not a blog, and not patent claim text.

See also: Research papers

reading pathPlatform

A sequenced route through Technology pages matched to an audience (buyer, builder, cryptographer, or ceremony-curious) so you learn the security-class distinction before diving into method or formal math.

See also: Where to start

reserveLedger

The ledger operation that allocates a fresh nonce with a timestamp and a time-to-live, marking it RESERVED.

RosarioSecurity

ENI6MA’s reference proof construction for Channel Zero–style authorization: a public ceremony whose transcript is designed so observers learn nothing usable about the hidden secret under stated axioms. The How Rosario works page walks the method step by step.

See also: How Rosario works · Axioms

security paradigmSecurity

A class of security design (computational hardness, device possession, succinct proof artifacts, post-quantum transport, or information-theoretic empty-channel authorization) compared by job and metric, not by vendor trophy.

See also: Compare security systems

session independenceSecurity

Fresh challenge geometry and entropy each round so transcripts do not correlate into a secret map across sessions. Watching many logins does not assemble φ when each ceremony is independent under the model.

See also: Why classical attacks fail · How Rosario works

session-bound no-replaySecurity

Axiom P4: every proof is bound to a freshness epoch (τ / nonce). A transcript from one session is not accepted in another. Operational latch, often earned in deployment via burn-before-validate.

See also: Axioms · Security

Shor's algorithmSecurity

A quantum algorithm that efficiently solves certain algebraic problems underlying RSA and elliptic-curve public-key systems on a large enough quantum computer. It changes hardness calendars for structured crypto, not a universal decoder for empty transcripts.

See also: Quantum myths vs real bounds

sixth primitiveSecurity

Name for a cognitive–cryptographic authorization primitive that sits beside encryption, hashes, MACs, signatures, and zero-knowledge proofs: prove knowledge now, for this request, without leaving a stealable credential on the channel. “Sixth” is ladder language, not an ISO taxonomy number.

Formal note: Ladder language for a presence/membership proof primitive beside the classical five (encrypt, hash, MAC, sign, ZK). Target posture: one-shot, request-bound evidence with transcript silence $I(X;O\mid C)=0$ under the model, not a portable SNARK artifact.

See also: A new kind of proof

synonym-channel uniformityMath

Axiom that, given the public challenge, every public response symbol is equiprobable. Flat histograms deny frequency analysis: counting symbols yields no compass toward the secret. Often labeled P2 in the formal pillar set.

See also: Uniform channel · Axioms

tau (τ)Protocol

The timestamp, in microseconds since the UNIX epoch, that the ledger sets when a nonce is reserved. It is bound into the proof and checked against a freshness window, so a captured envelope goes stale in seconds.

See also: Notation: τ · How Rosario works

uniform bearingsMath

Public labels chosen so each symbol is about equally likely. Flat frequency defeats classical letter-counting attacks: the channel can look busy (nonzero surface entropy) while still carrying zero mutual information about the secret under the model.

See also: A flat, unreadable channel

unstructured searchMath

Finding a marked item in a combinatorial space without algebraic shortcuts. Membership-only empty-channel designs often reduce the attacker to this model (Grover/BBBV bounds), where only leaf identifiers appear, not an invertible witness.

See also: Quantum myths vs real bounds

zero-knowledge proofSecurity

A cryptographic argument that a statement is true without revealing the secret witness that makes it true. Common constructions (ZK-SNARKs, STARKs) usually rest on computational hardness and often produce a portable proof artifact, useful for many jobs, different from an empty authorization channel.

See also: Why not all proofs are equal · Compare security systems

Notation

Public-safe symbol card for Math pages. Deep links use /technology/glossary#notation and #notation-<slug>.

ITS and computational entries both appear above; cross-read Computational bargain, Perfect secrecy, and Comparison matrix.