pods-topic-selection
GitHub用于判断数据管理项目应投稿至PODS、SIGMOD/VLDB等系统会议，还是ICDT、LICS/STOC等理论期刊或会议。依据贡献是定理还是系统实现进行路由决策。
Trigger Scenarios
Install
npx skills add brycewang-stanford/Awesome-Journal-Skills --skill pods-topic-selection -g -y
SKILL.md
Frontmatter
{
"name": "pods-topic-selection",
"description": "Use when deciding whether a data-management project belongs at ACM PODS (the database-theory symposium) or should be routed to SIGMOD\/VLDB\/ICDE (systems), ICDT (its sister theory venue), LICS\/ICALP\/STOC (pure theory), or a journal (TODS\/LMCS\/JACM), by contribution shape, the result-swap test, and the PODS-vs-ICDT community fit."
}
PODS Topic Selection
Decide the venue before drafting. PODS — the ACM SIGMOD/SIGACT Symposium on Principles of Database Systems — is the theoretical-foundations symposium held jointly with SIGMOD each year. Its papers are theorems about data management: models, query languages, complexity bounds, dichotomies, logic, and provably optimal algorithms. A technically strong paper whose real contribution is a faster system, a benchmark win, or an engineering artifact is respected and then rejected as out of scope — that paper is SIGMOD/VLDB/ICDE. PODS reviewers read for a provable, foundational statement about a cleanly defined model.
The routing question that matters most
The decisive question is rarely "is this about databases?" but "is the contribution a theorem or a system?" If the headline is a bound, a dichotomy, a semantics, an expressiveness result, or an algorithm with a matching lower bound, it is PODS-shaped. If the headline is throughput, latency, or accuracy on real workloads — even with clever theory inside — its home is a systems-DB flagship. PODS and SIGMOD share a week and a hallway but not a bar for acceptance.
Sibling-venue routing table
| Signal in your project | Better home | Why |
|---|---|---|
| A bound, dichotomy, semantics, or provably optimal algorithm in a data model | PODS | The database-theory symposium; results are theorems with proofs |
| A system, index, or optimizer evaluated on real workloads for performance | SIGMOD / VLDB / ICDE | The systems-DB flagships; evidence is measured, not proved |
| Database theory, but you want the EDBT/ICDT federation or missed the PODS cycle | ICDT | PODS's sister theory venue; overlapping community and reviewer pool, different calendar |
| Pure logic / finite model theory with no data-management payoff | LICS / ICALP / STOC / FOCS | Theory venues; PODS wants the data-management motivation to be central |
| A deep, long development beyond a 15-page symposium result | TODS / LMCS / JACM / VLDBJ | Journals with no symposium page ceiling; often the full-version home |
| Applied ML-for-data with empirical validation as the point | SIGMOD / VLDB / a ML venue | Not a PODS theorem; route to where the evidence is measured |
Contribution shapes PODS rewards
- A new semantics or framework for an ill-defined problem — consistent query answering, provenance semirings, a data model for uncertainty — stated precisely and studied for decidability/complexity.
- A complexity classification — a dichotomy (PTIME vs. hard), a fine-grained bound, a data-vs-combined-complexity separation over a query class.
- An expressiveness or logic result — which logical language captures an operation (composition, view definition, path querying), with matching upper and lower bounds.
- A provably optimal algorithm — a worst-case-optimal join, a constant-delay enumeration procedure, an MPC-round bound — the guarantee is a theorem, not a measured speedup.
- Foundations of emerging data problems — differential privacy for queries, learning-theoretic guarantees for data tasks, graph-query theory — where PODS supplies the rigor.
The result-swap and re-label tests
Two quick tests sharpen a borderline verdict:
- Result-swap test: if you replaced your specific algorithm or construction with a different one, would a theorem still remain (a bound, a classification, an impossibility)? If not, the artifact is the contribution and a systems venue fits better.
- Re-label test: could this paper be submitted to SIGMOD unchanged and read as native there? If its heart is a measured system with theory as garnish, route to SIGMOD/VLDB; if its heart is a proof, PODS is home. The mirror also holds for ICDT — if the paper is pure logic with no data-management question driving it, ICDT or LICS may fit better than PODS.
Maturity, without the ladder cliché
Fit is necessary but not sufficient. A conjecture with partial evidence but no proof is a workshop or a Gems-of-PODS talk, not a research paper; a theorem whose proof only handles a special case needs the general result or an honest scope; a sprawling development that cannot breathe in 15 pages plus appendix may belong in a journal first, with a PODS extended abstract of the core theorem. Submitting a half-proved result costs a full cycle even when the topic is perfect.
Cheap reconnaissance before committing
[Scope] scan the last two PODS proceedings (dblp db/conf/pods, PACMMOD PODS track) for your
subarea -> 3+ recent papers = a reviewer pool exists; 0 = mismatch or ICDT/LICS territory
[Citations] is your bibliography majority PODS/ICDT/LICS/TODS/JACM, or majority SIGMOD/VLDB?
-> majority systems venues => reviewers may read you as a systems paper; reframe the intro
[Calendar] PODS runs two cycles a year; compare the next PODS cycle with ICDT's and the systems
deadlines, and route to the nearest honest fit rather than idling
Decision procedure
[Contribution] theorem (bound/dichotomy/semantics/optimal algorithm) or measured system?
[Model] is there a cleanly defined data/query model the result is about?
[PODS vs ICDT] data-management theory with the SIGMOD community pull -> PODS; EDBT/ICDT federation
fit or calendar -> ICDT
[Systems check] performance is the headline -> SIGMOD/VLDB/ICDE, not PODS
[Verdict] PODS / ICDT / systems flagship / journal-first, with a one-line reason
Run this before the writing skills; a wrong venue decision wastes every later step. When the verdict
is PODS, continue with pods-workflow for the two-cycle calendar and pods-writing-style for the
paper shape.
Version History
- 9f86f09 Current 2026-07-19 17:17
