Agent Skills
› brycewang-stanford/Awesome-Journal-Skills
› itcs-review-process
itcs-review-process
GitHub建模ITCS会议审稿流程，强调无反驳阶段及轻量双盲机制。指导作者通过前置写作应对误解，侧重概念新颖性评估，明确PC决策权，帮助作者校准预期并优化投稿策略。
Trigger Scenarios
询问ITCS会议审稿规则
了解无反驳阶段的应对策略
咨询如何提升概念新颖性权重
查询ITCS投稿时间节点
Install
npx skills add brycewang-stanford/Awesome-Journal-Skills --skill itcs-review-process -g -y
SKILL.md
Frontmatter
{
"name": "itcs-review-process",
"description": "Use to model the ITCS reviewing pipeline — a program committee plus external reviewers, conceptual-novelty weighting, lightweight double-blind, no author-response\/rebuttal phase, and a single accept\/reject decision — and to understand where the (limited) author leverage actually is."
}
ITCS Review Process
Understanding how ITCS decides tells you where your effort matters — and at ITCS, almost all of it matters before the deadline, because there is no rebuttal, no revise-and-resubmit, and no second round. The submission is judged as it stands. This skill models the pipeline so you can calibrate expectations and pre-empt the objections you will never get to answer.
The pipeline
- Registration then submission. Title/abstract/authors/conflicts lock ~2 days before the paper (early September); the abstract drives PC bidding. Then the anonymized full PDF with complete proofs is uploaded to HotCRP.
- PC + external review. A program committee (order of a few dozen members) handles the load; PC members review directly or solicit external reviewers (subreferees) for expertise. The PC — not the external reviewers — owns the decisions.
- Reviewing on the novelty axis. Reviewers are briefed to weight conceptual contribution
— a new model, question, or connection — over incremental technical depth. "Is this a new
question worth asking, and is the model alive?" carries more weight than "is this the year's
hardest theorem?" (contrast STOC/FOCS; see
itcs-topic-selection). - PC decision. The committee reaches accept/reject. There is no Major Revision category and no author-response phase — the reviews are formed and the decision made without your input.
- Notification (~November), then camera-ready for LIPIcs.
Lightweight double-blind, precisely
- The submitted PDF omits author names, affiliations, and emails; the prose avoids identity-revealing self-reference.
- But it is lightweight: the reference list is not anonymized, and authors are encouraged to post a non-anonymous full version to arXiv/ECCC/ePrint. Reviewers are not forbidden from encountering the preprint. The blinding reduces first-impression bias; it does not attempt airtight anonymity.
- Some editions have handled this as a "reveal after initial review" middle ground; treat the
exact mechanism as cycle-volatile and read the current call (see
../../resources/official-source-map.md).
The no-rebuttal reality
This is the single biggest difference from many CS venues, and it reshapes strategy:
- You cannot correct a misunderstanding after the fact. If a reviewer misreads your model or doubts your novelty, the paper must have forestalled it in the text.
- The submission is the whole argument. Everything a depth venue's authors would save for a rebuttal — the "is it trivial?" defense, the novelty delta, the scope caveats — must be present at upload.
- See
itcs-author-responsefor how to operate when there is no response phase, and how the pre-emption work happens inside the paper.
Where the (limited) leverage is
| Stage | Leverage | How to use it |
|---|---|---|
| Before submission | High | Pre-empt the three objections (trivial? new? why-care?) in the paper; prove the anchoring result in full |
| Abstract/registration | Medium | A precise, honest abstract improves PC bidding and reviewer match |
| Preprint | Low-medium | A clear full version helps a reviewer who wants more detail (allowed under lightweight double-blind) |
| After reviews arrive | ~None | No rebuttal; reviews are usually returned with the decision |
| After reject | Medium (next venue) | Reviews are typically substantive; use them to sharpen a STOC/FOCS/SODA or journal resubmission |
What reviewers are looking for (calibrate to these)
- A genuinely new idea, checkably new against the ITCS/ICS back-catalog and the connected area.
- A model that is alive — an anchoring separation/possibility/impossibility result.
- Complete, correct proofs of every central claim (the appendix counts; the external preprint does not, on its own).
- Honest scope — a bold first foray with candid open problems is a feature here, not a weakness.
- A reason the community should care — a connection, a reframing, an application sketch.
Common misreadings to avoid
- Treating ITCS like STOC and optimizing for depth — the PC may find the paper strong but "not innovative enough."
- Assuming a rebuttal will let you fix a gap — there is none; the gap sinks the paper.
- Assuming the external full version will be read — the PC judges the submitted PDF.
- Anonymizing references or self-citations — that breaks the lightweight rule and can read as
evasive (see
itcs-related-work).
Output format
[ITCS review calibration]
[Novelty] is the new idea legible and checkably new to a briefed PC? yes/no
[Alive] anchoring result present? yes/no
[Proofs] all central claims fully proved IN the submitted PDF? yes/no
[Objections pre-empted] trivial? / new? / why-care? all answered in-text (no rebuttal to rely on)? yes/no
[Anonymity] names/affil/email removed, references NOT anonymized? yes/no
[Risk read] likely reviewer objection + where the paper already answers it (or must)
Version History
- 9f86f09 Current 2026-07-19 16:06
