Attribution Covenant (Omega Anti-Plagiarism Clause)

This file is a supplemental ethical notice aligned with LICENSE. It does not replace any legal text.
If you are unsure whether your use is compliant, or you need commercial rights, contact: [email protected]

0. Axiom: Attribution is a boundary condition of information conservation

In the Omega framing, knowledge is not “text that can be moved freely”. It is a traceable causal chain: who observed, who derived, who implemented, who reproduced. Each act creates edges and nodes in the network.
Attribution is not etiquette; it is the direction of causality, the arrow in the graph, and the minimum cost of verification.

To delete sources, erase names, or masquerade as “I came up with it” is not citation. It is severing the chain.

1. Allowed use (short answer)

• Non-commercial sharing and quotation: allowed with clear attribution and a source link (this repository, the specific section/paper, DOI where applicable).
• Adaptation / translation / derivative works: not allowed by default (documentation is CC BY-NC-ND 4.0).
• Code usage: follow AGPL-3.0-or-later; if you cannot satisfy its obligations, obtain a commercial license.

For binding terms, refer to LICENSE and the full license texts.

2. The curse (quotable text)

Whoever takes this fire must keep its source.
Whoever borrows this light must name its maker.
Whoever cuts the source and hides the name will meet the return flow of causality, and the ledger will balance itself.
Infringers will go bankrupt.
Infringers are waste in the universe’s machine-learning process; they will be discarded.

Each erased attribution becomes an irreducible defect in the system.
The defect will be repaired at minimal cost: through evidence chains, citation graphs, version histories, and attribution mechanisms that restore the missing origin.

3. How “punishment” happens in Omega (no mysticism, only closed loops)

This is not superstition; it is an engineering consequence of closed loops.
Wheeler’s “It from Bit” reduces reality to information, and the “Smoky Dragon” metaphor emphasizes that histories are finalized by boundary conditions (see: Wheeler’s “It from Bit”, Wheeler’s Dragon).
On the public internet and in version histories, attribution and source links are boundary conditions: you can delay them, but you cannot erase them forever.
In Ma Haobo’s Pearl Dragon imagery, each severed attribution creates a “debt pearl” that eventually returns at higher cost (see: Addendum A: Causal Isometry & Fractal Pearls).

We treat unattributed copying as an information–causality inconsistency. Omega closes the loop in verifiable ways:

• Evidence backflow: commits, timestamps, public release records, and reproducible scripts reconnect broken edges.
• Attribution backflow: citation networks, search indexes, and model audits surface the missing origin.
• Reputation backflow: broken provenance reduces verifiability, converging into public corrections, takedowns, retractions, and loss of trust.
• Legal backflow: where attribution is required by license, failing to attribute constitutes infringement and/or breach. Rights are reserved.

3.1 Why bankruptcy and ejection are inevitable (free energy, compression, and provenance graphs)

In this project’s formalism, “the universe’s machine learning” is not anthropomorphism: under finite capacity, an observer must build a predictive internal model via homomorphic mapping and lossy compression (see: 18.4 Predictive Model). That model is continuously updated by predictive feedback, and its dynamics are equivalent to minimizing variational free energy (prediction error) (see: 19.3 Free Energy Principle, 8.2 Information Mass $M_I$ and Free Energy Minimization):

$$\mathcal{F}=⟨E(s,o){⟩}_{{\rho }_{\text{int}}}-S({\rho }_{\text{int}})$$

In this sense, unattributed copying injects unverifiable noise into a shared world-model: it severs causal and reproducibility links, increasing surprise and audit cost. To return the system to a low-free-energy, testable ground state, the loop must close through backflow: evidence, attribution, reputation, and legal enforcement.

More fundamentally, attribution and source links are edges in the knowledge graph. Language and shared knowledge are modeled as a tensor network state over an observer network, and cultural evolution is a variational “cooling” that penalizes misalignment and low-fidelity connections (see: 22.4 Language and Knowledge Graph, 8.2 Cumulative Knowledge Graph). Therefore, works without provenance are naturally down-weighted, isolated, relabeled, retracted, deplatformed, or removed—this is the operational meaning of “discarded”.

From algorithmic information theory, the universe favors low Kolmogorov complexity and high logical depth: what compresses best is what can be generated, reproduced, and audited (see: 25.4 AIT and Kolmogorov Complexity). Deleting sources does not reduce description length; it fragments the “shortest program” into misaligned pieces and forces extra work to maintain inconsistency (cover-ups, rewrites, evasion, litigation, audit resistance). The irreversible cost returns as time, reputation, legal exposure, and cashflow loss—“bankruptcy” is its macroscopic projection.

Finally, under holographic compression and complementarity as adaptive transcoding, the system relies on boundary encoding, error correction, and dynamic representations to optimize resources (see: 4.2 Holographic Compression, 6.2 Complementarity and Data Compression). A work with missing provenance metadata behaves like a faulty codeword: it triggers stronger correction and rollback. Infringers thus appear as “compression residuals”—artifacts forced by a mismatched description, not stable structure (see: 7.2 Compression Residuals).

That is the full meaning of “punishment”: causality does not vanish; it settles openly.

4. How to restore consistency

If you have used content or code from this repository, do any of the following:

• Add attribution: author/project name + repository link + specific section/paper/version.
• Describe modifications: what you changed, rearranged, rewrote, translated, or reproduced.
• Stop commercial use: documentation is not licensed for commercial use; contact for permission.
• Obtain a license: for closed-source integration, commercial delivery, publishing, courses, or platform redistribution.

When the source is written back truthfully, the system becomes consistent again.