socc-workflow
GitHub用于规划ACM SoCC会议投稿流程的技能，涵盖从选题、实验到双轮次提交、反驳及最终展示的完整时间线管理。支持基于截止日期的反向计划，强调云系统论文的实验锁定与双轮次不可重投规则，确保符合会议规范。
触发场景
安装
npx skills add brycewang-stanford/Awesome-Journal-Skills --skill socc-workflow -g -y
SKILL.md
Frontmatter
{
"name": "socc-workflow",
"description": "Use when planning an ACM SoCC project timeline from venue fit through the two-round abstract+paper deadlines, the rebuttal, notification, camera-ready, and presentation, with backward-planning offsets for a cloud-systems measurement paper and honest handling of the two-rounds-per-year cadence and the no-cross-round-resubmission rule."
}
SoCC Workflow
Use this as the project-management skill for a SoCC submission. Replace every date with the current official timetable and work backwards from the round's abstract-registration cutoff — which precedes the paper deadline by about a week. SoCC — the ACM Symposium on Cloud Computing, jointly sponsored by SIGMOD and SIGOPS — runs two review rounds per year, so the first planning decision is which round you are genuinely ready for.
SoCC is a conference with no standing editor-in-chief; its rotating leadership is the per-edition General Chair(s) and Program Committee Co-Chairs appointed under SIGMOD and SIGOPS (SoCC 2026 PC Co-Chairs reported as Eric Lo and Ivan Beschastnikh, verified 2026-07-09 via committee-page rendering; confirm on the live page). The cost model is conference registration; at least one author presents.
Milestones
- Venue fit: confirm SoCC over OSDI/NSDI/EuroSys/ATC and over SIGMOD/VLDB
(
socc-topic-selection), and confirm the contribution spans the systems-and-data intersection. - Evidence lock: freeze the system/mechanism, the testbed or trace, the workloads, the baselines, and the measurement pipeline.
- Round choice: pick Round 1 or Round 2 knowing a Round-1 reject cannot be resubmitted to Round 2 the same year.
- Abstract registration: submit real title, abstract, authors, subject tags, and conflicts by the round's abstract cutoff.
- Submission: upload the anonymized acmart PDF and (where possible) the artifact link by the round's paper deadline (AoE).
- Rebuttal: during the round's author-response window, answer reviewers on the axes they raised.
- Notification: accept / reject for that round; archive reviews, rebuttal, and every version.
- Acceptance: prepare the camera-ready, pursue ACM artifact badges if the edition offers evaluation, register, and present in Singapore (SoCC 2026).
Backward plan from a round's paper deadline
| Weeks out (heuristic) | Cloud-systems milestone |
|---|---|
| 10+ | System/mechanism feature-frozen; testbed or trace and baselines chosen |
| 8 | Experiments running; measurement harness logs throughput, latency, tail, and cost |
| 6 | Analysis done; cost-latency and tail results computed; baselines tuned with a documented budget |
| 4 | Full draft in acmart sigconf; artifact (trace-replay/testbed scripts) assembled and anonymized |
| 3 | Internal mock review by a systems reader and a data/measurement reader |
| 2 | Tail and cost results hardened; related-work delta sharpened; 12-page budget met |
| 1 | Dual-anonymous sweep on PDF and artifact (traces, cluster names, system names) |
| 0 | Abstract registered by its earlier cutoff, then paper + artifact link on HotCRP |
These offsets are planning heuristics only — anchor every one to the current Important Dates / HotCRP deadlines page, never to a previous cycle's calendar.
Two-rounds-per-year: the honest calendar reality
SoCC's twin-round model is a genuine advantage and a trap:
- Advantage: there are two shots per calendar year at the same venue (a winter/February round and a summer/July round for SoCC 2026), so a paper that misses one round is only months from the next.
- Trap: a paper rejected in Round 1 may not be resubmitted to Round 2 the same year, so a rushed Round-1 submission does not buy a discounted retry — it costs you the year at SoCC. Submit to the round you can make strong, not merely the nearer one.
- As of 2026-07-09, the live target is SoCC 2026 Round 2 (paper 14 Jul 2026); Round 1 (13 Feb 2026) has passed.
Failure modes by stage
- Evidence still moving at week 4 forces last-minute measurement nobody has audited — the classic soundness reject for a systems paper.
- Simulation-only at submission invites the "does it hold on a real testbed?" reject from both sponsoring communities.
- Reporting only the mean — omitting tail latency and cost — reads as not understanding what a cloud paper is judged on.
- Leaving the anonymized trace/artifact to the final week is how a cluster name or provider hint leaks identity under dual-anonymous review.
- Treating Round 1 as a free draft and getting rejected forfeits the whole SoCC year.
Coordination notes
- Assign one owner for the measurement harness and reproducibility package and another for the dual-anonymous sweep; shared ownership is how both slip.
- Archive the exact submitted PDF and artifact — the rebuttal must quote the paper precisely and stay anonymous.
Output format
[Current stage] idea / evidence / round-choice / submitted / rebuttal / accepted
[Target round] round 1 / round 2, with the next official deadline and source
[Critical path] <three tasks that determine readiness>
[Risk register] <page budget / anonymity / simulation-vs-testbed / tail+cost evidence / artifact>
[Owner map] <task -> person or role>
版本历史
- 9f86f09 当前 2026-07-19 17:42
