vldb-review-process
GitHub解析 PVLDB 期刊式月度审稿流程，涵盖六周周期、单次修改机会、单盲评审及接受/修改/拒绝结果。用于设定投稿预期、解读审稿意见并指导后续修改或重投策略。
触发场景
安装
npx skills add brycewang-stanford/Awesome-Journal-Skills --skill vldb-review-process -g -y
SKILL.md
Frontmatter
{
"name": "vldb-review-process",
"description": "Use when reasoning about how PVLDB reviewing actually works, covering the rolling monthly pipeline, the roughly six-week review turnaround, the accept\/revise\/reject outcome space, the one-shot revision's odds, single-blind dynamics, the volume review board, and how a VLDB decision differs from batch-deadline venues."
}
VLDB Review Process
Use this to set expectations after (or before) a PVLDB submission. The process is a journal pipeline wearing a conference badge: a standing review board for the volume, decisions on a monthly cadence, and a revision outcome with teeth. Details below were verified 2026-07-08 for Vols 19/20; the live guidelines page governs.
The pipeline at a glance
25th: abstract -> 1st: paper (5PM PT) -> ~15th next month: reviews + verdict
|
accept ---------------- 3 outcomes ---------------- reject
|
revise (one-shot)
|
up to 3 months -> revised paper -> same reviewers
|
accept or reject
Roughly six weeks from deadline to decision is the structural promise of the rolling model — competitive with any CS venue — and the price is that there is no rebuttal stage: your first chance to talk back is the revision, if offered.
What each outcome means here
| Outcome | Practical meaning | Your move |
|---|---|---|
| Accept | Publication in an upcoming monthly issue; conference slot at the matching VLDB | Camera-ready discipline (vldb-camera-ready) |
| Revise | Reviewers believe a specific, listed set of fixes could flip them | Requirement ledger + three-month plan (vldb-author-response) |
| Reject | The paper as framed did not convince this panel | Rework before any re-entry; check current re-submission rules (待核实 whether an embargo applies) |
A revise verdict is genuinely positive signal — the board spends its future reviewing time only on papers it considers rescuable.
Who is reading
- The volume operates with Editors-in-Chief (who double as the conference PC chairs — Vol 20: Boehm and Hose per secondary sources) atop a standing review board assembled for the volume.
- Expect builder-reviewers: people who wrote storage engines, optimizers, and the baselines you measured. Implementation hand-waving and untuned competitors are caught by lived experience, not by checklists.
- Review is single-blind: reviewers see your names. Track records cut both ways — a known group gets less benefit of the doubt for overclaiming, not more.
Reading reviews on the monthly clock
- The verdict letter's required-changes list is the only text that matters for a revise; reviewer musings outside it are advisory.
- For a reject, extract the decision reason (novelty vs. evidence vs. fit) before touching the draft — rolling venues tempt authors into resubmitting cosmetically-edited papers at machine-gun pace, and review boards notice.
- Timing tactics around "easier months" are folklore; the board is standing and the bar does not oscillate. Pick the month your evidence is ready.
Contrast with batch venues
At a batch-deadline conference a rejection costs a year and a rebuttal is a 480-word ritual; at PVLDB a rejection costs a rework cycle, and the revision is a real second review with new evidence allowed. Optimize accordingly: at this venue, holding a submission until the weakest experiment is fixed is cheap, and burning the revision on a preventable flaw is expensive.
Output format
[Stage] pre-submission / under review / revise / rejected / accepted
[Clock position] <deadline used, expected decision date>
[Outcome analysis] <required-changes list or decision reason>
[Panel model] <likely reviewer profile for this topic>
[Next move] <hold, revise per ledger, rework, or camera-ready>
[Rules to reverify] <re-submission terms, current escalation channel>
版本历史
- 9f86f09 当前 2026-07-19 17:53
