Popl topic selection

POPL Topic Selection

POPL's full name is the routing test: Principles of Programming Languages. The venue's center of gravity is a general truth about languages — a semantics, a type discipline, a program logic, a proof technique — established by definition and theorem. Both theoretical and experimental papers are welcome under the posted scope, but an experimental POPL paper still orbits a principle. (Scope wording per the POPL series pages, read 2026-07-08.)

Three questions before committing

1. What sentence do you most want reviewers to believe? If it starts "we prove" or "we give a semantics/logic/type system such that ..." — POPL. If it starts "our compiler/analyzer achieves ..." — PLDI. If it starts "developers using our language ..." — OOPSLA.
2. Could a skeptic check the claim without running anything? POPL's decisive evidence is re-derivable: proofs, mechanizations, counterexample-freedom by construction. If the truth of your claim lives in measurements, route toward the implementation venues.
3. Does the idea outlive your artifact? POPL rewards machinery others can instantiate on their languages. A theorem inseparable from one codebase is a tool paper wearing a theorem.

Routing table

Project signature	Better venue	Reason
New program logic + soundness proof + small case studies	POPL	Principle first, evidence supporting
Compiler optimization with 1.4x speedups, proofs optional	PLDI	Benchmark-carried claim, implementor jury
Language design evaluated by user study or corpus mining	OOPSLA	Broader empirical evidence culture
Typed FP abstraction, elegant library, paradigm-native	ICFP	The functional-programming home crowd
Pure model/automata/logic result, thin language connection	LICS / CSL / FSCD	Logic-side siblings want it more
Decision procedure, model checker, solver improvement	CAV / TACAS	Verification-algorithm reviewers
Proof-engineering technique, assistant infrastructure	CPP / ITP	Mechanization-first audience (CPP co-locates with POPL)
Solid but regional-scale PL result, or needs journal length	ESOP / TOPLAS / JFP / LMCS	Right bar, right format

All of POPL, PLDI, OOPSLA, and ICFP publish in PACMPL, so the archival venue is identical — choose by reviewer community and evidence shape, never by deadline proximity.

POPL-fit smells, both directions

• Good signs: the abstract states a theorem; the motivating example breaks an existing system's metatheory; reviewers from three different PL subareas would each recognize the machinery.
• Warning signs: the "formal" section restates the implementation in Greek letters; the theorem is routine and the contribution is engineering; the paper needs performance numbers to matter.

Fit test (fill honestly):
  Central claim: ..............................................
  Decisive evidence:      [ ] proof   [ ] mechanization   [ ] measurements   [ ] study
  Who must believe it:    [ ] semanticists  [ ] implementors  [ ] tool users
  Survives artifact deletion?  [ ] yes  [ ] no
  => POPL only if column one dominates and the answer is "yes".

Output format

[POPL fit] strong / plausible with reframing / route elsewhere
[Central claim] <the one-sentence version>
[Evidence shape] <proof / mechanization / empirical mix>
[Reframe path] <what to foreground if staying at POPL>
[Alternative] <venue + one-line reason, if routing away>