Editorial constitution

Trust should be visible in the record.

The journal is being built around a simple obligation: readers must be able to understand the argument, assess the evidence, and carry an insight into practice without losing its limits or provenance.

The record standard

These are launch requirements, not an aspirational badge. Public claims will not imply DOI registration, indexing, or preservation until each service is verified.

One structured record

JATS XML is the canonical article record; accessible HTML, PDF, metadata, and agent representations are derived from the same version.

JATS 1.4

Authenticated identity

ORCID identifiers will be collected through authentication—not typed into a box—and contribution roles will use the CRediT taxonomy.

ORCID criteria

Two kinds of expertise

Peer-reviewed work is evaluated by both a research or method specialist and an experienced education-practice leader. Editors remain responsible for decisions.

NISO peer-review terminology

Corrections are first-class

Version history, corrections, expressions of concern, and retractions propagate across the article, downloads, discovery metadata, and Paper Agent index.

Crossmark

Access is the whole journey

WCAG 2.2 AA is the baseline for reading, submitting, reviewing, editing, and proofing—not only for public article pages.

WCAG 2.2

Preservation, not just backup

Versioned deposits, stable identifiers, and continuous preservation are part of publication. A server backup is not a scholarly archive.

DOAJ guidance

Human accountability

What automation may never decide.

Acceptance or rejection

Editors—not scores, language models, or integrity tools—make and explain every decision.

Reviewer fitness

Automation may organize declared expertise; it cannot determine independence or replace conflict review.

Research integrity

Signals prompt proportionate human inquiry. A flag is not a finding of misconduct.

Authorial responsibility

Automated tools cannot be authors. Material use must be disclosed and every claim remains the authors’ responsibility.

Paper Agent architecture

The agent is derived. The article remains the authority.

We adapt Paper2Agent’s strongest ideas—source snapshots, provenance manifests, review gates, verified tool boundaries, and independent checks—without treating an executable agent as the scholarly record itself.

01

Version lock

Every answer identifies the exact article version and becomes stale when that version is corrected or withdrawn.

02

Passage evidence

Substantive answers link to stable paragraphs, tables, figures, notes, or review-history records.

03

Honest refusal

When the selected record does not support an answer, the agent says so plainly.

04

Execution boundary

Executable supplements run separately in hardened, disposable workers—never in the public journal application or confidential review workflow.

Open foundation, explicit boundary

Paper2Agent is used under the MIT License, © 2025 Jiacheng Miao. Its research-code execution workflow is not deployed inside this journal prototype.

View source

Accessibility & openness

No account is required to read.

Public articles, review status, disclosures, corrections, and accessible full text will remain open. Sign-in is reserved for private records and attributed actions such as submissions, reviews, proofs, and saved reading.