Journal wrapper: launch decision and mechanics, throughput-gated (journal phase 3)

[larnsce]

Phase 3 of the journal strategy (#91). This is a decision issue with a hard gate, not a commitment. Evidence base: dev/journal-2026-07/research-data-journals.md on main.

Launch gate (all three, measured — decision point ~2028)

1. Sustained reviewed throughput ≥ 20 datasets/year for two consecutive quarters (from the phase-2 pilot metrics, Editorial layer: published review criteria, two-reviewer open review, editor recruitment, pilot (journal phase 2) #95). Rationale: journals publishing <20–25 articles/year look fragile or dead to indexers and authors; the Journal of Open Humanities Data manages ~76/year with an entire discipline behind it.
2. ≥ 3 active handling editors who are not Lars (volunteer burnout is the documented killer of small journals; JOSS survives on volunteer editors + bot automation + NumFOCUS/Sloan backing).
3. A funded managing-editor fraction (ORD Phase 2 data-stewards line or successor funding).

If the gate fails: stop at phase 2. A certified, reviewed catalogue with DOIs is a respectable permanent end state; nothing from phases 1–2 is wasted. Revisit yearly.

If launched — mechanics

• Form: short data descriptors (2–4 pages), substantially generated from the package via generate_paper() (generate_paper(): draft a data descriptor from package contents (preprint route into Google Scholar) #94); diamond OA, no fees ever (aligned with Gates 2025 no-APC policy). The dataset DOI (Zenodo/DataCite) and the paper DOI (Crossref: $1/article + ~$275/yr membership) are distinct and cross-linked via relatedIdentifiers.
• Infrastructure: Open Journals stack (JOSS's editorialbot is open source, already reused by rOpenSci) or plain Quarto + GitHub. JOSS's audited marginal cost: ~$3–6/paper. Archival arrangement (Portico ~$250/yr or CLOCKSS) early — indexers check for it.
• Conventional shell around open review (the JOSS lesson): JOSS has no impact factor after a decade because Scopus/WoS treat its checklist review as unconventional, while ESSD (open interactive review, conventional editorial form) carries IF 13.6. Keep GitHub review as the mechanism; add the form indexers evaluate: ISSN at launch; named editorial board with affiliations and international + Global South representation (grown from Editorial layer: published review criteria, two-reviewer open review, editor recruitment, pilot (journal phase 2) #95's editors); formally documented two-reviewer policy; COPE-aligned ethics pages; fixed publication cadence (continuous publication with quarterly issue rollups); deliberate publisher-of-record decision (ETH library imprint vs association vs foundation — affects perceived independence and continuity).
• Indexing sequence: DOAJ within year 1 (near-free, achievable, the anti-predatory signal); Google Scholar via proper article metadata (Highwire citation_* tags + PDFs — closes the visibility gap in Discoverability: Zenodo datasets are structurally invisible in Google Scholar — channel strategy for FAIR publishing #93); apply to Scopus and WoS ESCI only after ~2 years of regular publishing (failed applications trigger embargo periods; Scopus real acceptance of suggested titles is <20%; since the 2023 JCR every ESCI journal receives an impact factor, making ESCI the fastest IF route). Realistic first IF: year 3–5 post-launch, if at all.
• Positioning: "the venue where WASH data becomes a first-class, peer-reviewed, citable publication — free for authors and readers." The impact factor is never the pitch until it exists (openwashdata journal: three-phase strategy (submission policy → editorial layer → journal wrapper) #91 kill criterion 4 — promising it early is the predatory-journal tell and burns the credibility everything depends on).

Strategic hedge

By the time a first IF can arrive (~2030), data-citation metrics (DataCite Commons, Make Data Count, Data Citation Corpus) and funder mandates may have shifted how data credit is counted. The journal's value must stand without the IF: Crossref DOI + Scholar visibility + peer-review quality signal + the maintenance contract (phase 1, #92) — a maintained dataset with a responsive contact is something no incumbent data journal offers. Treat the IF as an option that may arrive, not the promise.