thedixitjain OOPSLA Submission
OOPSLA accepts new papers into either of two yearly rounds feeding the PACMPL
OOPSLA volume; for the 2026 volume the rounds closed October 10, 2025 and
March 17, 2026, both 11:59:59 PM AoE and explicitly firm, via
oopsla26.hotcrp.com. Before running this audit, confirm which round is
actually open on the current SPLASH track page — as of 2026-07-08 both 2026
rounds had closed and the 2027 Round 1 call was not yet posted (待核实).
The three questions that precede any checklist
- Which round? A round is a strategic choice, not just a date. Filing in
the earlier round leaves the later round of the same volume available for
a Major-Revision resubmission; filing in the later round pushes any Major
Revision into next year's volume. See
oopsla-workflowfor the calendar math. - New paper or invited revision? A Major-Revision invitation must come back as a revision in the designated round, answering the reviewers' expectation list — never re-filed as a fresh submission. Mixing these up forfeits reviewer continuity.
- Is the evidence finished? The two-round model gives reviewers a revision lever; an evaluation you plan to "complete during review" invites exactly that lever to be pulled against you.
Format contract (2026 volume)
| Rule | Value | Consequence of violating |
|---|---|---|
| Class options | acmsmall,screen,review,anonymous on acmart |
Wrong template reads as non-compliant at triage |
| Initial submission cap | 23 pages, excluding references, required statements, appendices | Over-cap papers risk rejection without review |
| Invited-revision cap | 25 pages, same exclusions | Extra two pages exist to answer reviewers |
| Data-Availability Statement | Required, placed before the references, not counted | Missing statement is an avoidable flag |
| Anonymity | Double-anonymous, multi-stage | Identity leaks void the review contract |
Anonymity sweep beyond the obvious
- Tool and prototype names that are already public: rename or cite in the third person, and scrub GitHub organization strings from listings.
- The Data-Availability Statement itself must stay anonymous — point to an anonymized archive, not a lab server whose hostname names your group.
- PDF metadata (
pdfinfoshows author fields LaTeX quietly embeds), embedded figure paths, and acknowledgment stubs left from a prior journal draft. - Self-citation grammar: "we extended X [12]" de-anonymizes; "this work builds on X [12]" does not.
Concurrent-submission and prior-work posture
Declare any overlapping manuscript honestly. The dangerous case at OOPSLA is the sibling-venue loop: a rejected PLDI or POPL version resubmitted with the same evaluation weaknesses — reviewer pools overlap, and the two-round model means the same people may see the paper twice within one year. Fix what the last venue's reviews established before re-entering the family.
Filing-day sequence
T-3 days Freeze claims; regenerate all numbers from the artifact
T-2 days Anonymity sweep (text, metadata, statement, supplement)
T-1 day HotCRP record: title/abstract match PDF, topics, conflicts,
all authors listed with institutions in the form only
T-0 Upload PDF + supplement; re-download and open both;
confirm timestamp shows AoE margin, not local-time optimism
Deadlines being firm means the plan needs no heroics on the final night; if the paper is not assembled at T-3, target the next round instead — with two rounds a year, the cost of waiting is months, not a year.
Output format
[Round] <target round + deadline, from the live track page>
[Filing type] new submission / invited revision
[Format] pass / fail per contract row above
[Anonymity] clean / leaks found: <list>
[Statement] Data-Availability Statement present and anonymous: yes/no
[Verdict] file now / fix list / wait for next round
Source: brycewang-stanford/Awesome-Journal-Skills → OOPSLA-Skills/skills/oopsla-submission/SKILL.md