Grand Challenge Reader Guide

A reading protocol for locating the object, obstruction, support route, claim boundary, and next executable move.

Start from the reader's question

Use the shortest path that answers the question you have.

Question	First page
What is active now?	Current Work
What has crossed a checked or published boundary?	Results and Exemplars
How does the programme move claims?	Programme Atlas
Which standing domains and archives exist?	Mathematical Estate

Do not infer current priority from the domain catalogue. Do not infer theorem status from activity, publication, merge state, or visual prominence.

Five questions govern serious reading

1. What is the object?

Identify the exact mathematical object before following the machinery around it.

Ask:

What data specifies one instance?
What ambient structure or model class is assumed?
Which source or construction defines the object?
What counts as the same object under the current formulation?

2. What is the obstruction?

Locate the smallest known reason the next claim does not already follow.

The obstruction can be a missing lemma, failed hypothesis, singular case, source gap, model-class mismatch, unavailable invariant, undecidability barrier, or failed certificate replay.

A general statement that the problem is difficult is not an obstruction.

3. What is the strongest supported claim?

Separate the claim from the programme's larger ambition.

Determine whether support is:

exploratory;
audited;
exactly computed;
certificate-replayed;
formally proved;
a continuum proof;
a negative result.

Then identify the scope, assumptions, and limitations attached to that support.

4. What crossed a trusted boundary, and what did not?

Check the semantic bridge between the intended human statement and the formal or computational object.

A trusted checker can establish only the statement that it actually checks. Source concordance, model-class restrictions, translation assumptions, and omitted hypotheses remain separate obligations unless they are also discharged.

5. What is the next bounded move?

A useful next step names its input, action, output, and completion test.

Examples include:

replay one theorem dependency;
construct one exact certificate;
search one bounded counterexample class;
prove one restricted lemma;
reconstruct one missing source correspondence;
measure one declared engineering response map.

"Continue research" is not a next move.

Read through the correct lens

New contributor

Reconstruct the landscape before proposing a route. Begin with the object, source boundary, first example, and current obstruction.

Campaign author

Check whether the local obligation is sharp enough to attack. A Work Package should identify dependencies, completion conditions, and claim firewalls.

Formalization or certificate lead

Compare the intended statement with the encoded statement. Record assumptions, imports, model restrictions, trusted environment, and semantic-bridge debt.

Skeptical reviewer

Ask what evidence would downgrade the claim. Check whether the support class, current revision, source identity, and retained exclusions match the public summary.

Anatomy of a serious artifact

A reader-facing research artifact should expose these parts in a recoverable order:

Status. Current claim and lifecycle state.
Plain object. The mathematical object before machinery.
Exact obstruction. The smallest mechanism blocking the naive route.
Working model. A hand-checkable example, diagram, or bounded fixture.
Restricted claim. The exact local statement under investigation.
Theorem-spine location. Dependencies and role in the larger route.
Support route. Proof, computation, replay, certificate, or negative result.
Debt and boundary. What remains open, imported, conditional, or unaudited.
First executable step. One finite continuation action.

This is the programme's nine-move exposition pattern. An artifact can compress it, but it should not silently omit a stage that matters to interpretation.

Reading a live front

For a live front, use this order:

Read its entry in Current Work.
Open the live tracker named there.
Recover the current object and obstruction.
Follow the protected predecessor or evidence record.
Identify the exact claim firewall.
Confirm that the next move has a completion test.
Treat the tracker as navigation; use protected repository records for authority.

Reading a bounded result

For a result or exemplar:

Start at Results and Exemplars.
Identify the support class.
Follow the exact verification, certificate, publication, or archive record.
Check retained exclusions.
Distinguish mathematical support from artifact lifecycle or publication status.
If a new programme starts from the result, verify that it does not silently broaden the earlier claim.

Reading a standing domain

The Mathematical Estate is a scope catalogue, not a priority list.

For a domain:

Open its public landing page.
Follow the canonical repository entry.
Read the claim boundary before the historical narrative.
Use current-work links only when that domain has a live active front.
Follow Work Packages by dependency, not merely by date.
Inspect a proof, replay, or certificate before treating a result as promoted.

Visuals

A diagram should teach a structure that prose is slower at teaching.

Before relying on a visual, ask:

What mathematical relation is literal?
What placement, scale, geometry, or styling is schematic?
What source or data generated the visual?
Does the visual summarize evidence, or merely explain it?

A visual does not create proof authority. A schematic must not be read as a literal mathematical model unless its contract says which relations are exact.

Final test

After reading a serious artifact, a reader should be able to state:

We studied this object. The obstruction was this. The strongest supported claim is this. This part crossed a trusted boundary. This part remains open. The next bounded move is this.

If those six statements cannot be recovered, continue to the underlying record before relying on the summary.