brycewang-stanford OOPSLA Related Work
OOPSLA sits inside a family that shares authors and reviewers: POPL, PLDI, ICFP publish in the same PACMPL journal, ECOOP is the adjacent European OO venue, Onward! shares the SPLASH building, and ICSE/FSE/ISSTA overlap on empirical software questions — ISSTA was literally co-located with SPLASH 2026. The related-work section is therefore written for people who may have authored the cited lines, and who will meet your paper again if a revision round occurs. Two failure modes dominate: venue misattribution and delta-free citation lists.
Citation form in the journal era
Since OOPSLA became a PACMPL track, its papers are journal articles and must be cited as such:
@article{belyakova2020worldage,
author = {Belyakova, Julia and Chung, Benjamin and Gelinas, Jack
and Nash, Jameson and Tate, Ross and Vitek, Jan},
title = {World Age in Julia: Optimizing Method Dispatch
in the Presence of Eval},
journal = {Proc. ACM Program. Lang.},
volume = {4},
number = {OOPSLA},
articleno = {207},
year = {2020},
doi = {10.1145/3428275}
}
Recent volumes split into OOPSLA1 and OOPSLA2 issues — preserve the issue
name (number = {OOPSLA1}) exactly as the ACM DL records it. Citing a PACMPL
article as "In OOPSLA '24" is the kind of sloppiness family reviewers notice
in their own papers' entries.
Verification discipline
Every venue claim gets checked, not remembered:
| Check | Tool | What it catches |
|---|---|---|
dblp key (conf/oopsla/ or journals/pacmpl/) |
dblp search | PLDI/POPL/ECOOP results misfiled as OOPSLA (see the cut list in resources/exemplars/library.md) |
| DOI resolves to claimed venue | ACM DL | Preprint-vs-published divergence; wrong-year citations |
| Issue OOPSLA1 vs OOPSLA2 | ACM DL table of contents | Half-year attribution errors in recent work |
| Sibling series coverage | Per-venue dblp streams | The "we searched OOPSLA only" blind spot |
Deltas, not inventories
The family's standard is one technical difference per cited line of work, in the sentence where the citation appears:
- Weak: "Ownership types have been widely studied [3,7,12,19]."
- Strong: "[3] and [7] assume whole-program annotation; [12] recovers partial soundness but reports blame at the failure site rather than the boundary; we prove boundary blame is the weakest sound choice (§3)."
For concurrent work surfacing in the other round of the same volume — a genuinely OOPSLA-specific situation — cite it, mark it concurrent, and state the relationship without claiming priority; reviewer continuity means the committee likely knows both papers.
Coverage strategy
- Sweep the last three volumes of PACMPL Issues OOPSLA1/OOPSLA2, plus the matching POPL/PLDI/ICFP issues, before declaring novelty.
- Add ECOOP and the relevant SE venue (ICSE/FSE/ISSTA) whenever the claim is empirical or tooling-facing; add TOPLAS/JFP for journal-depth lines.
- Onward! papers are citable idea-stage precedents: acknowledging one costs nothing and omitting one your reviewer wrote costs a lot.
- Keep a claims-to-prior-work matrix in the repo; regenerate it at each round, because seven months pass between rounds and the family publishes continuously.
Output format
[Attribution audit] entries checked on dblp/DL: n/n, errors: <list>
[Citation form] PACMPL journal form used for post-2017 OOPSLA: yes/no
[Delta coverage] cited lines with explicit deltas: n/m + worst offenders
[Family sweep] OOPSLA/POPL/PLDI/ICFP/ECOOP/SE volumes covered
[Concurrent work] same-volume-round items flagged and handled