DeSci Protocol Framework
Core decentralized science coordination framework for on-chain research proposals, peer review, funding, and publication.
ZIP-600: DeSci Protocol Framework
Abstract
This ZIP defines the foundational Decentralized Science (DeSci) protocol for the Zoo ecosystem. It establishes on-chain primitives for the full research lifecycle: proposal submission, peer review coordination, funding allocation, milestone tracking, data publication, and reproducibility verification. The protocol creates a permissionless, transparent, and incentive-aligned system for scientific research that preserves academic rigor while removing gatekeepers from the funding and publication pipeline.
Motivation
Traditional scientific research suffers from well-documented structural failures:
- Funding bottlenecks: Grant agencies reject 75-85% of proposals, and review cycles take 6-18 months. Early-career researchers and non-institutional scientists are systematically disadvantaged.
- Publication bias: Journals favor positive results, creating a replication crisis. Negative results and null findings are rarely published despite their scientific value.
- Opaque review: Peer review is closed, slow, and inconsistent. Reviewers receive no compensation or reputation credit for quality work.
- Data silos: Research data is locked behind institutional paywalls and proprietary formats, hindering reproducibility and meta-analysis.
- Misaligned incentives: Researchers optimize for publication count and impact factor rather than scientific truth and reproducibility.
Zoo Labs Foundation, as a 501(c)(3) focused on open AI research and conservation science, requires a protocol-level framework that aligns incentives toward reproducible, open, and impactful research. This ZIP provides that framework, building on the ethical foundations of ZIP-540 (Research Ethics & Data Governance) and integrating with the broader Zoo governance system defined in ZIP-0 (Ecosystem Architecture).
Cross-Ecosystem Context
This proposal draws on governance patterns from:
- HIP-0066 (Hanzo Data Governance): Data classification, access control, and lifecycle management primitives.
- HIP-0045 (Hanzo Documentation Framework): Structured metadata and versioning for technical artifacts.
- LP-8800 through LP-8805 (Lux DAO Governance Platform): On-chain proposal lifecycle, voting mechanics, and treasury integration.
Specification
1. Research Proposal Lifecycle
A research proposal is an on-chain object with the following state machine:
DRAFT --> SUBMITTED --> UNDER_REVIEW --> FUNDED --> ACTIVE --> COMPLETED
| | | |
v v v v
WITHDRAWN REJECTED CANCELLED FAILED
1.1 Proposal Schema
{
"proposalId": "bytes32",
"version": "uint16",
"title": "string",
"abstract": "string",
"ipfsCid": "string",
"principalInvestigator": "address",
"coInvestigators": ["address"],
"institution": "string",
"category": "enum(CONSERVATION, AI, BIODIVERSITY, CLIMATE, SOCIAL)",
"fundingRequested": "uint256",
"currency": "address",
"milestones": [
{
"description": "string",
"deliverables": ["string"],
"fundingPortion": "uint16",
"deadline": "uint64"
}
],
"ethicsReviewStatus": "enum(EXEMPT, EXPEDITED, FULL)",
"dataManagementPlan": "string",
"openAccessCommitment": "bool",
"createdAt": "uint64",
"updatedAt": "uint64"
}
1.2 Submission Requirements
| Field | Requirement |
|---|---|
| Full proposal document | IPFS-pinned PDF/Markdown with methodology, budget, timeline |
| Ethics classification | Self-assessed per ZIP-540 categories (Exempt/Expedited/Full) |
| Data management plan | Per ZIP-540 FAIR/CARE principles |
| Open access commitment | All outputs CC-BY or CC0 unless restricted per ZIP-540 |
| Milestone breakdown | Minimum 2 milestones; each with measurable deliverables |
| Conflict of interest | Disclosed affiliations and funding sources |
1.3 Proposal Bond
Submitters deposit a bond of 100 ZOO (configurable by governance) to prevent spam. The bond is returned upon completion or if the proposal reaches UNDER_REVIEW regardless of outcome. Bonds are slashed only for proposals found to violate ZIP-540 ethics standards.
2. Peer Review Coordination
Peer review is managed on-chain with off-chain review content stored on IPFS.
2.1 Reviewer Assignment
1. Proposal enters SUBMITTED state
2. Review coordinator (elected per ZIP-603) assigns 3-5 reviewers
3. Reviewers are selected from the Reviewer Registry based on:
- Domain expertise tags
- Reputation score (see Section 2.3)
- Absence of conflict of interest (see ZIP-603)
- Availability and recent workload
4. Reviewers accept or decline within 7 days
5. Declined slots are reassigned automatically
2.2 Review Schema
{
"reviewId": "bytes32",
"proposalId": "bytes32",
"reviewer": "address",
"ipfsCid": "string",
"scores": {
"scientificMerit": "uint8(1-10)",
"methodology": "uint8(1-10)",
"feasibility": "uint8(1-10)",
"impact": "uint8(1-10)",
"ethics": "uint8(1-10)"
},
"recommendation": "enum(FUND, REVISE, REJECT)",
"confidenceLevel": "uint8(1-5)",
"submittedAt": "uint64"
}
2.3 Reviewer Reputation
Reviewers accumulate reputation through a staking and scoring mechanism:
| Action | Reputation Effect |
|---|---|
| Complete review on time | +10 base points |
| Review quality score > 8/10 | +5 bonus points |
| Review cited by other reviewers | +3 per citation |
| Review overturned by appeal | -15 points |
| Missed deadline | -10 points |
| Conflict of interest violation | -50 points, temporary suspension |
Reputation decays at 5% per quarter to incentivize continued participation.
3. Funding Mechanism
3.1 Funding Sources
| Source | Mechanism |
|---|---|
| Zoo Treasury | Direct allocation via ZIP-603 governance |
| Quadratic Funding Rounds | Community-matched donations per ZIP-603 |
| Earmarked Grants | Donor-restricted funds per ZIP-0 Section 11 |
| Retroactive Public Goods | Post-completion rewards for high-impact research |
3.2 Milestone-Based Disbursement
Funds are held in escrow and released per milestone:
Milestone submitted by PI
|
v
Milestone reviewed by 2+ reviewers (7-day window)
|
v
Milestone approved (majority vote)
|
v
Funds released to PI address
Failed milestones trigger a 30-day remediation period. Two consecutive failed milestones move the proposal to FAILED state and remaining funds return to the treasury.
3.3 Escrow Contract Interface
interface IResearchEscrow {
function depositFunds(bytes32 proposalId) external payable;
function releaseMilestone(bytes32 proposalId, uint16 milestoneIndex) external;
function failMilestone(bytes32 proposalId, uint16 milestoneIndex, string calldata reason) external;
function cancelProposal(bytes32 proposalId) external;
function getProposalBalance(bytes32 proposalId) external view returns (uint256);
}
4. Publication and Data Registry
4.1 Publication Record
All funded research must produce a publication record on-chain:
| Field | Description |
|---|---|
| proposalId | Link to original proposal |
| ipfsCid | Full paper/report on IPFS |
| datasetCids | Associated datasets on IPFS |
| codeCid | Analysis code repository hash |
| license | CC-BY-4.0, CC0, or approved alternative |
| peerReviewCids | Post-publication review records |
| reproducibilityStatus | VERIFIED, UNVERIFIED, FAILED |
4.2 Reproducibility Verification
Independent researchers may submit reproducibility reports:
1. Claim a publication for reproducibility verification
2. Execute methodology using published data and code
3. Submit reproducibility report with results comparison
4. Report reviewed by original reviewers + 1 independent
5. Publication record updated with reproducibility status
Successful reproducibility verification earns both the original PI and the verifier reputation and token rewards.
5. Smart Contract Architecture
ResearchRegistry (upgradeable proxy)
|
+-- ProposalManager
| Manages proposal lifecycle and state transitions
|
+-- ReviewCoordinator
| Assigns reviewers, collects reviews, computes scores
|
+-- FundingEscrow
| Holds funds, releases per milestone approval
|
+-- PublicationRegistry
| Records outputs, tracks reproducibility
|
+-- ReputationTracker
Manages reviewer and PI reputation scores
All contracts are deployed behind UUPS proxies with a 48-hour timelock on upgrades, controlled by the Zoo governance multisig.
6. Protocol Parameters
| Parameter | Default | Governance-Adjustable |
|---|---|---|
| Proposal bond | 100 ZOO | Yes |
| Review period | 30 days | Yes |
| Reviewers per proposal | 3 | Yes (min 2, max 7) |
| Milestone review window | 7 days | Yes |
| Remediation period | 30 days | Yes |
| Reputation decay rate | 5% / quarter | Yes |
| Minimum reputation to review | 50 points | Yes |
Rationale
- On-chain state machine: Embedding the proposal lifecycle on-chain provides an immutable audit trail and enables trustless milestone disbursement. Off-chain alternatives lack the transparency required for a 501(c)(3) operating in public.
- IPFS for content: Full proposals, reviews, and publications are too large for on-chain storage. IPFS provides content-addressed, decentralized storage with the on-chain CID serving as a tamper-proof reference.
- Reputation over identity: The reviewer reputation system avoids requiring real-world identity verification (which conflicts with reviewer anonymity goals in ZIP-604) while still providing sybil resistance through staking and decay.
- Milestone-based funding: Lump-sum grants create moral hazard. Milestone-based disbursement aligns incentives and gives the community early exit if research goes off track.
- Retroactive funding: Including retroactive public goods funding acknowledges that the most impactful research is often only recognized after completion.
Security Considerations
- Sybil attacks on review: Mitigated by minimum reputation thresholds, stake requirements, and conflict-of-interest detection (ZIP-603). A single entity controlling multiple reviewer identities would need to accumulate reputation independently for each.
- Proposal spam: The bond mechanism prices out low-effort submissions. Governance can adjust the bond amount in response to spam volume.
- Escrow risk: Funds in escrow are at smart contract risk. Contracts must undergo formal verification and at least two independent audits before deployment. Emergency pause functionality is included.
- IPFS availability: Content pinning is required by the protocol. Zoo Labs operates pinning infrastructure and requires funded projects to use at least two independent pinning services.
- Oracle manipulation: Milestone approval is human-driven (reviewer votes), not oracle-dependent, reducing attack surface. Score aggregation uses median rather than mean to resist outlier manipulation.
- Governance capture: Protocol parameter changes require a 48-hour timelock and supermajority vote per ZIP-603, preventing rapid parameter manipulation.
References
- ZIP-0: Zoo Ecosystem Architecture & Framework
- ZIP-540: Research Ethics & Data Governance
- ZIP-560: Evidence Locker Index
- ZIP-603: Research DAO Governance
- ZIP-604: Decentralized Peer Review
- HIP-0066: Hanzo Data Governance
- HIP-0045: Hanzo Documentation Framework
- LP-8800 through LP-8805: Lux DAO Governance Platform
- DeSci Foundation. "Decentralized Science Manifesto." 2022.
- Buterin, V. "Retroactive Public Goods Funding." 2021.
Changelog
| Version | Date | Changes |
|---|---|---|
| 1.0 | 2025-01-15 | Initial draft |
Copyright
Copyright and related rights waived via CC0.