Research Review

Keep a record.
Work on a claim.

Start a local journal, connect observations and keep corrections. When a statement is ready for review, prepare a claim for your participant client.

Start or open a journal

Work stays in this tab until you download it: no upload, browser storage, wallet or signature. Open a journal (up to 8 MiB) or workflow packet (up to 16 MiB). Hash checks establish internal consistency, not authorship or scientific truth.

Or open a journal or workflow packet

JavaScript is required to validate and open a research file.

The interactive reader needs JavaScript. Download the fictional journal and inspect it with the local journal tools. No imported records have been validated or displayed here.

Download to keep the complete journal. Closing this tab discards changes that you have not downloaded.

Add a record or correction

Each addition keeps the earlier entries intact. A correction is a new record linked to an earlier version.

Name this record What do you want to record?

Attributed to Record type

Correct an earlier record · optional

Evidence, dates and connections

When it happened · optional DOI · optional

Evidence reference · optional Describe the evidence Source link File fingerprint (SHA-256) · optional

Connect to an earlier contribution · optional

This new record has the relationship below to the earlier contribution.

This new record… …this earlier contribution

Adds to the full journal in this tab. Nothing is submitted to the chain.

Write a record, then add it to the journal.

Choose a contribution

Record & evidence

Claim & review

Browsing keeps this draft. Using the selected record replaces it; download existing work first.

Choose the statement and checks you want reviewed. Editing this draft leaves the original journal record unchanged.

Select a contribution to prepare a packet.

The download contains the complete journal, including entries beyond a historical cutoff. Check it before sharing. The client submits only your chosen claim, reasoning and provenance hashes.

Use your participant client

From a verified source checkout, inspect the downloaded file, then submit with your existing home and verified client. Replace the two absolute paths below. Open the returned packet here to continue.

python3 tools/research-review/workflow.py inspect --packet ./zerone-research-workflow.json
python3 tools/research-review/workflow.py submit --packet ./zerone-research-workflow.json --home /absolute/participant-home --client /absolute/verified-package/shared-claims.py
python3 tools/research-review/workflow.py refresh --packet ./zerone-research-workflow.json --home /absolute/participant-home --client /absolute/verified-package/shared-claims.py

Recover an interrupted submission

Use refresh first. If the packet remains started without a receipt, deliberately resume through the existing client’s recovery guards. Do not submit it again.

python3 tools/research-review/workflow.py resume --packet ./zerone-research-workflow.json --home /absolute/participant-home --client /absolute/verified-package/shared-claims.py

Reviewers choose and sign their own reasons and verdicts. This page never holds a key or signs a transaction.

Participant setup and review guide → · Workflow commands and journal tools →

Chain observations have their own block heights and times, separate from the journal cutoff. An inconclusive review stays inconclusive; corrections and payments appear only as recorded.

Native history requires a successful submission receipt. Saved observations are dated endpoint state, not fresh reads or proofs of truth.

Journal history and relationships

In this collection by entry Order contributions by

Contributions

Relationships

Recorded relations to the selected contribution. A highlighted path suggests what to inspect; it does not establish falsity.

Visible relation types

Concerns in this snapshot

Attributed leads for review, not misconduct verdicts.

All assessments in this snapshot

Every author and scope is retained. Later assessments do not erase disagreement.

Compare complete journals

Choose a second journal export. Comparison is independent of the timeline cutoff; it does not choose a canonical or trustworthy branch.

Second local export · up to 8 MiB

Open a primary collection first. JavaScript is required for comparison.

Try fictional branch A and branch B. Open A above, then compare B here. Neither depicts an actual external reviewer.

About the records and their limits

The reader cutoff changes which records you inspect. The editor’s earlier-record choices and additions use the complete current journal; downloads keep the complete journal. Earlier events discovered later remain absent from earlier reader snapshots. Unknown dates stay unknown.

For a correction, use “Add a record or correction” and choose the earlier record. Both versions remain in the journal. Earlier assessments do not transfer to the correction; a native claim needs its own explicit preparation and submission.

Possible reassessment follows supports from evidence to claim, requires from prerequisite to dependent, and uses-input from input to activity. Shared input is labelled separately. Citation, inspiration and refinement do not propagate the highlight.

Attributions are declared labels. The fictional example depicts no real research, people or on-chain state. Evidence links open only when you choose them; they are not verified here.

Download a dated research checkpoint · AdaGrad experiment · Journal tools and format documentation · Download the fictional collection

Keep a record.Work on a claim.