naacl-workflow
GitHub用于规划NAACL研究项目日历，处理其跳过年份的特殊节奏。涵盖ARR周期选择、反向推导数据标注时间线，并提供在提交、等待或转投其他会议时的决策树指南。
Trigger Scenarios
Install
npx skills add brycewang-stanford/Awesome-Journal-Skills --skill naacl-workflow -g -y
SKILL.md
Frontmatter
{
"name": "naacl-workflow",
"description": "Use when planning a research project's calendar toward NAACL — handling the chapter's skipped-year rhythm, choosing an ARR cycle when the edition's feeding cycles are not yet announced, backward-planning data and annotation work, and keeping a decision tree for commit-versus-wait-versus-reroute moments."
}
NAACL Workflow
Project management for NAACL differs from every annual venue in one blunt way: the conference might not exist next year. The chapter historically skips editions when the ACL main meeting is hosted in the Americas — no NAACL 2023, no NAACL 2026 — so a NAACL plan is really a plan for entering the *ACL Rolling Review machine at a moment that leaves a NAACL commitment open, with named fallbacks for when the calendar shifts under you.
The calendar reality (as of 2026-07-08)
- Most recent editions: NAACL 2024 (Mexico City, June 16-21, 2024) and NAACL 2025 (Albuquerque, April 29 - May 4, 2025).
- Next edition: NAACL 2027 — site registered, but dates, venue, feeding cycles, and commitment deadline all 待核实 at planning time.
- Reference shape from 2025: ARR upload mid-October, reviews mid-December, commitment around the turn of the year, final versions early February, conference late April. A spring NAACL implies an autumn ARR deadline — plan provisionally on that shape, verify the moment the CFP lands.
Provisional backward plan for a spring NAACL (verify every date):
T-0 conference (spring)
T-11w final versions due
T-16w commitment deadline <- reviews must exist by here
T-26w ARR cycle submission <- the real deadline
T-30w evidence freeze: data, annotation, main runs complete
T-38w annotation campaign starts (if collecting new data)
T-46w pilot studies + guideline drafting
Milestones with owners
| Milestone | Owner | Exit criterion |
|---|---|---|
| Venue decision review | Whole team | NAACL still the target? (rerun quarterly — see decision tree) |
| Data/annotation freeze | Data lead | Agreement scores computed; splits published internally |
| Main-results freeze | Experiments lead | Ledger complete; variance reported |
| Paper freeze | Writing lead | Content pages self-contained; Limitations drafted |
| Compliance pass | One named person | Checklist, anonymity, template, citations verified |
| ARR upload | Corresponding author | Confirmation email archived |
| Response window | All authors on call | Every review answered per triage |
| Commitment decision | Senior author | Decision logged with rationale |
Annotation campaigns deserve the earliest slot they can get. Recruiting annotators for Guaraní or Haitian Creole is not a two-week task, and community partnerships — common for Americas-language projects — move at partnership speed, not deadline speed.
The commit-versus-wait decision tree
Reviews arrive; now the genuinely NAACL-specific decision:
- Strong package, NAACL window open → commit to NAACL.
- Strong package, no NAACL window soon → commit to the nearest fitting sibling (EMNLP, EACL, AACL, or ACL) rather than letting reviews age a year; NAACL loyalty is not worth a stale paper.
- Fixable weaknesses → revise and resubmit to a later ARR cycle with a change summary, timed so the new reviews land before the NAACL commitment deadline.
- Field moving fast under the paper → prefer earlier venues over waiting; a skipped-year gap is long enough for your baselines to become historical artifacts.
Log the decision and the date; teams that skip the log relitigate the same choice monthly.
Monitoring routine between editions
Between the decision to target NAACL and the edition's CFP, run a light monthly watch instead of a panicked scramble later:
- ARR dates page: has any cycle acquired a NAACL assignment?
- Chapter site and the edition placeholder (
{year}.naacl.org): dates, venue, or organizing-committee announcements. - The chapter's announcement channels and the ACL Member Portal, where cycle-to-conference mappings are often posted first.
- On any hit: update the backward plan the same week, and rerun the decision tree with the real dates.
Assign the watch to one named person; a "we all sort of watch" arrangement reliably discovers CFPs three weeks late, which on a ten-week cycle is fatal.
Standing risks on a NAACL path
- Announcement risk: feeding cycles for a new edition can be announced with little slack; subscribe to the chapter's channels and recheck the ARR dates page monthly once an edition is expected.
- Pool-depth risk: post-gap commitment pools are deep; a borderline package should strengthen or reroute, not hope.
- Single-point compliance risk: desk rejects (template, Limitations, checklist) burn the whole cycle — ten weeks lost, not ten days.
- Deadline convention risk: all ARR clocks run AoE (UTC-12); a team spread across the Americas' time zones still gets exactly one cutoff.
Output format
[Target edition] <NAACL year, or "unannounced — provisional plan">
[Cycle chosen] <ARR month/year + verification date>
[Backward plan] <milestone -> date -> owner>
[Decision-tree state] commit / wait / revise / reroute (+ rationale)
[Top calendar risk] <one line>
Version History
- 9f86f09 Current 2026-07-19 17:05
