Docs / Proposing a topic

Proposing a topic

A topic is a big question of a field, often loosely posed, that open problems sit under. The first topics came with the seed; anyone with an ORCID iD may propose more. A proposal is a signed record, like a problem: it names its field and carries the topic's title, a summary, why it matters and, optionally, a review that covers it.

From signature to decision

  1. Sign. "Propose a topic" on a field's page opens the form; you sign the record with your ORCID iD. At most 5 topics a day per account. You can also propose one together with a new problem (below).
  2. Screening. With your consent, an AI screening checks that it is a topic of physics, not too narrow, not a duplicate of an existing topic and not spam (AI at PubPhys). A confident pass publishes it; anything else, or no consent, holds it for an editor of the field, who releases it or rejects it at screening. While it is screened or held, only you and the field's editors see it.
  3. Proposed. Once published, the topic is public, labelled "Proposed, not yet reviewed", and problems may be signed under it.
  4. Decision. An editor of the field, or the chief editor, accepts or rejects it with signed reasons. Nobody decides on a topic they proposed, revised, or have a problem waiting under (who may decide). The first acceptance of a topic is told to the followers of its field (follows).

With a new problem

On the page of your new problem, the topic list offers "Propose a new topic" when no topic fits. Its fields open under the list. If you let AI read your file, it may have drafted them already, when none of the field's topics fitted: the draft is only a proposal, which you edit before you sign, and the topic's page then says "drafted with AI help" (site data, not part of the signed record; a topic record declares no AI). The topic and the problem are signed in one ORCID sign-in, as one batch: both are recorded or neither. The pair counts once toward the daily limit of topics and once toward that of problems.

The topic is screened first, and the problem waits for it. Once the topic is published, the problem goes through its own screening and is published as unreviewed under the proposed topic. If the topic is held for an editor, the problem is held with it and goes on once the topic is released or rejected. If the topic is rejected at screening or removed, the problem is still screened (by AI only under the consent you gave for the problem, otherwise by an editor), but stays hidden, on every page, in search and in the API, until an editor moves it to an accepted topic (below).

Problems under a proposed topic

Only a topic's proposer may sign new problems under it before an editor accepts it: the new-problem form lists your own published proposals after the field's accepted topics. Such problems are public, unreviewed, while the topic is; an editor decides the topic first, and only then the problems under it. Authors of problems waiting under a topic do not decide it. If the topic is rejected, its problems stay public under it, labelled; if it is removed, they are hidden; either way an editor moves them to an accepted topic (below).

Revising a topic

Until an editor decides it, its proposer may sign a new revision with "Edit this topic"; it goes through the same screening and replaces the waiting revision once published. After acceptance, editors of the field revise it: an editor's revision is screened, published as waiting, and decided by another editor, while the accepted revision stays displayed. Every revision is a new signed record; earlier ones stay in the log.

Appeals

The signer of a topic's revision may appeal its rejection, or its rejection at screening, within 30 days, while topics are accepted, from "Editors' decisions" on the topic's page (appeals). An editor with no part in the case decides the appeal. A grant restores the revision as it was before the decision: a rejected revision waits for an editor again, a revision rejected at screening is held again. Once the topic has changed since the decision (another revision was accepted or published), the appeal becomes moot: sign a new revision instead. The authors of problems under the topic are not parties to its appeals. The site's own refusal of a revision whose base changed while it was held is not appealable on its own.

Rejected and removed topics

A rejected proposal stays public, labelled with the editor's reasons, and its proposer can appeal the decision (above). The problems under a rejected topic, or under one rejected at screening, are listed for the field's editors, who move each one to an accepted topic with a signed decision.

Editors remove a topic only for copyright, personal data, abuse or spam. A removed topic leaves a tombstone; its text is withheld and its record's hash stays in the log. Its problems are hidden until an editor moves them to another topic.