跳到主要内容
知仓学习社ZHICANG

popl-supplementary

Use when splitting a POPL paper across the 25-page body, the proof appendix, and anonymous supplementary material — deciding which proofs and auxili…

不碰外部(只输出文字)无严重或高危命中brycewang-stanford/Awesome-Journal-Skills

它会碰到什么

扫了多少1 个文本文件,4 KB
它会碰到什么不碰外部(只输出文字)
命中总数0 处
命中统计严重 0 · 高 0 · 中 0 · 低 0

这一栏是扫描器报的事实,不是结论。命中多不等于有毒(安全工具、规则库、示例脚本本来就会包含危险写法),命中少也不等于干净。它和你手上的凭据、文件、网络有什么关系,需要你自己看。

技能内容

POPL Supplementary Material

A POPL submission is really three documents: 25 pages of text (bibliography

excluded) that must carry the whole argument, an appendix of full proofs and

auxiliary definitions, and — usually — an anonymized proof development or prototype.

Reviewers are typically expected to judge the paper from the body alone, so the split

is an argumentation decision, not a storage decision. Format and anonymity rules per

the POPL 2027 call, read 2026-07-08; confirm the current cycle's supplement wording

before uploading.

What lives where

| Content | Body (25 pp) | Appendix | Anonymous artifact |

|---|---|---|---|

| Main definitions, typing/semantics rules actually discussed | yes | mirrored in full | formalized |

| Main theorem statements + proof sketches | yes | full proofs | checked statements |

| Auxiliary lemmas, weakening/substitution boilerplate | no | yes | yes |

| Full figure of every judgment (all rules) | representative rules only | complete figure | source of truth |

| Extended examples, failed design alternatives | one motivating example | yes | test files |

| Proof scripts, build instructions | no | no | yes, with README |

Two disciplines make the split safe:

  • The body must stand alone. A reviewer who never opens the appendix should

still believe the theorem plausible from the sketch: state the invariant, the hard

case, and why it goes through. "Proof in appendix" after an unexplained claim

reads as a gap.

  • Sketch and proof must not disagree. The classic incident: the body's sketch

describes induction on typing derivations while the appendix inducts on evaluation

steps because the proof changed. Reviewers who notice stop trusting both.

Anonymizing a proof development

Full double-blind covers everything you upload. Proof repositories are leaky:

_CoqProject paths with usernames, lakefile package names matching a public

GitHub project, author headers auto-inserted by editors, and .git directories with

full commit history. Build the archive from an export, never from a working tree:

git archive --format=tar.gz -o /tmp/supp.tar.gz HEAD          # no .git, no untracked junk
tar tzf /tmp/supp.tar.gz | grep -iE '\.git|/home/|users/|TODO|AUTHORS' && echo LEAK
grep -rInE '(Copyright|Author|@[a-z]+\.(edu|org|fr|de))' \
  --include='*.v' --include='*.lean' --include='*.agda' extracted/ | head

Also rename the development if its public name is googleable to your group, and

strip institutional CI configuration.

Version-lock the trio

  • Tag the exact commit that generated the submitted PDF, appendix, and archive; the

author response will need to quote them line-precisely months later.

  • If the appendix is a separate PDF, give it the same section numbering scheme as

the body so "App. C.2" resolves unambiguously.

  • Late theorem renumbering must propagate to the correspondence table and README —

do it with a script, not by hand, in deadline week.

Output format

[Split audit] <claims whose evidence sits only outside the body>
[Sketch-proof consistency] <mismatches found>
[Anonymity scan] clean / leaks listed
[Version lock] <tag/commit for PDF + appendix + archive>
[Upload set] <files, sizes, formats for HotCRP>

想直接用这个技能?

本站把开放许可(MIT / Apache 等)的技能按仓库打包整理到网盘,点一下转存到你自己的网盘,不用一个个从 GitHub 拉。许可未声明的技能只给原始仓库链接,不打包。