pods-review-process
GitHub解释ACM PODS会议的审稿流程，涵盖双盲评审、多周期机制、两轮审阅及48小时反驳期。重点说明接受、拒绝及带导师监督的修订决策含义，并对比SIGMOD与ICDT的流程差异，指导作者针对理论证明的正确性与完整性进行策略性投稿。
Trigger Scenarios
Install
npx skills add brycewang-stanford/Awesome-Journal-Skills --skill pods-review-process -g -y
SKILL.md
Frontmatter
{
"name": "pods-review-process",
"description": "Use when reasoning about how an ACM PODS submission is evaluated, covering lightweight double-anonymous review, the multi-cycle-per-year calendar, the two reviewing rounds within a cycle, the 48-hour rebuttal, the accept\/reject\/revision decision with a shepherded revision, PACMMOD-track publication, and how PODS differs from SIGMOD's rounds and ICDT's process."
}
PODS Review Process
Model the pipeline before interpreting any single review. PODS review is done by database theoreticians who will read your proofs, and the process is multi-cycle with a revision round: a paper is not simply accepted or rejected on the first read — a revision decision (minor or major) with a shepherd is a first-class outcome. The most consequential mental shift for authors arriving from a plain accept/reject conference is that the revision is a real, shepherded second round inside the same cycle.
Process model
- Submission and review run on EasyChair with lightweight double-anonymity: author identities are hidden from reviewers, self-citations are third-person, and a conflict-of-interest list is declared at submission.
- Each cycle has two reviewing rounds to accommodate revisions. Reviewers assess the correctness and completeness of the proofs, the significance and novelty of the result, the precision of the model, and the tightness of the bounds.
- Authors get a short rebuttal window (about 48 hours, a few thousand characters) to correct factual misreadings before decisions.
- First decisions are accept / reject / revision. A revision (minor or major) invites a revised version within a set window; a shepherd judges whether the revision is satisfactory, and only then is the paper accepted.
- Accepted papers publish in the PACMMOD PODS track and are invited for presentation at the PODS symposium.
Reading a decision against the categories
| Decision | What it means | Author move |
|---|---|---|
| Accept | The result and proofs hold; at most cosmetic changes | Camera-ready + arXiv full version; do not reopen scope |
| Revision (minor) | A fixable gap: a clarified proof step, a stated assumption, a tightened bound | Implement precisely within the window; satisfy the shepherd |
| Revision (major) | A larger but plausibly closeable gap: a missing case, an unproven direction | Treat as a real second round; complete it or the shepherd declines |
| Reject | Structural: a wrong proof, an un-closable gap, or out-of-scope for a theory venue | Fix fully and route to a later cycle/ICDT/journal — no immediate resubmission |
The strategic reading: design the submission so its weakest point is repairable in a revision (a proof step to clarify, an assumption to state, a case to add) rather than structural (a main theorem that is false or unproven). The revision round rewards papers whose gaps are closeable.
How PODS differs from its siblings
- vs. SIGMOD/VLDB/ICDE: those systems flagships judge measured performance; PODS judges proofs. Even co-located with SIGMOD, PODS shares neither the acceptance bar nor the evidence type — a benchmark cannot rescue a missing proof here.
- vs. ICDT: ICDT is PODS's sister theory venue in the EDBT/ICDT federation with an overlapping community; the reviewing cultures are close, but the calendars and proceedings differ. Do not assume a shared deadline or template.
- vs. a once-a-year venue: PODS's multiple cycles mean a reject is only a season from another chance — but the resubmission embargo (roughly a year for rejected work) means a premature submission can still cost more than one cycle.
Who reads you
Expect theory reviewers who check the mathematics. They will read the key proofs, look in the appendix for the deferred ones, and catch a circular lemma, an unjustified "it is easy to see," or a lower bound that does not match the claimed upper bound. Because PODS spans logic, complexity, and algorithms, a paper is matched to reviewers in its subarea — a hand-waved proof is caught, not skimmed.
Where author leverage actually exists
[Before submission] topic tags + a precise abstract -> reviewer match (large lever)
[Rebuttal] correct factual misreadings and pointer errors in ~48 hours; not a place for new proofs
[Revision] the strongest lever: close the identified gap, satisfy the shepherd, resubmit in the window
[After reject] no appeal; fix fully, then a later cycle / ICDT / journal (mind the resubmission embargo)
The rebuttal moves a paper when it shows a reviewer misread a definition or missed an existing appendix proof; it does not move a paper by arguing taste or promising a proof you have not written. In a revision, an unaddressed required item is what turns the shepherded round into a reject.
Reading a review packet
Weight reviews before answering. A review that engages your theorem statements and points to a specific proof step read the paper closely and will re-check the revision — that reviewer or the shepherd is your path to acceptance if the fix holds. A review that only discusses significance has left correctness to the others; answer each reviewer on the axis they raised (correctness, tightness, novelty, modeling).
Misreadings to avoid
- Treating a revision as a guaranteed accept — the shepherd's second read is real; complete the required items, do not do half.
- Using the rebuttal to add new proofs — reviewers cannot verify unwritten mathematics in 48 hours; reserve substantive fixes for the revision.
- Assuming co-location with SIGMOD means a systems-friendly bar — PODS judges proofs, full stop.
- Projecting last cycle's calendar — cycle count and dates are decided per edition.
Output format
[Process stage] pre-submission / awaiting reviews / rebuttal / revision / final / accepted
[Decision category] accept / revision (minor|major) / reject, with the criterion driving it
[Criterion map] each review point -> correctness | tightness | significance | modeling | clarity
[Leverage plan] the next-stage action that can actually change the outcome (rebuttal fix / revision item)
[Forbidden moves] identity leak / new unverifiable proof in the rebuttal / ignoring a required revision item
Version History
- 9f86f09 Current 2026-07-19 17:16
