uist-review-process
GitHub解析UIST会议审稿流程，涵盖匿名评审、反驳机制及PC会议决策。提供2026年时间表，指导作者理解系统评审心理（如视频优先、新颖性评估），并教授如何分类回复审稿意见以优化结果或规划后续路径。
Trigger Scenarios
Install
npx skills add brycewang-stanford/Awesome-Journal-Skills --skill uist-review-process -g -y
SKILL.md
Frontmatter
{
"name": "uist-review-process",
"description": "Use when interpreting the UIST review pipeline — anonymous PC-plus-external review of papers and videos, the rebuttal's role, PC-meeting decisions, conditional acceptance mechanics, reading systems-reviewer psychology, and choosing the demo\/poster fallback or resubmission path after a rejection."
}
UIST Review Process
Understanding how UIST decides is mostly understanding who decides: a program committee of people who build interactive systems, supported by external reviewers drawn from the same community, converging in discussion and a PC meeting. The venue is single-track and relatively intimate — the person reviewing your haptics paper probably built a haptics device. The pipeline facts below are the 2026 cycle's (verified 2026-07-08); exact committee mechanics beyond the published dates are 待核实 each year.
The 2026 pipeline
| Stage | 2026 timing | What happens to your paper |
|---|---|---|
| Submission closes | Mar 31 | PDF + video + supplement enter PCS |
| Review period | April-May | PC members and externals review paper and video |
| Rebuttal | May 28 - Jun 5 | Authors respond in ≤ 5,000 characters |
| Discussion + PC meeting | June | Recommendations formed from submission, rebuttal, reviews, and reviewer/PC discussion |
| Notification | Jun 27 | Accept (conditional), or reject |
| Camera-ready | Jul 24 | Conditions from the rebuttal must be delivered |
Two structural notes: the decision explicitly weighs the rebuttal and the
inter-reviewer discussion, so the rebuttal is a real instrument (see
uist-author-response); and acceptances are conditional — the notification is a
contract whose terms you wrote yourself in the rebuttal.
How systems reviewers actually read
- Video first, skeptically. The initial question is "does this thing exist and work as claimed?" A missing video for a highly dynamic system starts the review from doubt.
- Novelty as capability delta. Reviewers pattern-match against the systems
canon; the related-work table is where they check whether you know it (see
uist-related-work). - Implementation as competence signal. Vague engineering prose reads as either concealment or shallow work; specific parameters and honest failure boundaries read as craft.
- Evaluation-matching, not evaluation-maximizing. A ritual study annoys
reviewers more than a well-argued demonstration portfolio (see
uist-experiments). - Enthusiasm is a real variable. UIST reviewers reward inventiveness; a technically modest but genuinely delightful idea can outscore a heavy but dull one. Write and film for the builder's grin.
Reading your reviews
Sort every reviewer point into one of four bins before reacting:
FACTUAL ERROR reviewer misread a mechanism/number -> correct, citing section
EVIDENCE GAP claim outruns measurement -> concede scope or add data
NOVELTY CHALLENGE "system X already does this" -> capability-delta answer
TASTE "I would have built it differently" -> acknowledge, do not litigate
The bins predict rebuttal leverage: factual errors and novelty challenges are winnable in 5,000 characters; evidence gaps are winnable only if the evidence already exists; taste is never winnable and costs characters.
Discussion dynamics in a small committee
Between rebuttal and notification, your paper's fate is a conversation you cannot attend. What is knowable about that phase shapes what you should have done earlier:
- Reviews converge through discussion; a champion reviewer arguing from your
rebuttal's specifics is the mechanism by which mixed reviews become accepts.
The rebuttal's job is to arm that champion (see
uist-author-response). - Expertise is uneven by design — a paper spanning fabrication and ML gets one deep expert per half; the review that "missed the point" often belongs to the other half's expert, and answering it teaches the whole committee.
- Single-track scale means PC members see a broad slice of the year's submissions; "we saw three of these this cycle" is a real dynamic, and a crisp capability delta is the defense.
- Scores move at the meeting. Borderline papers with concrete, deliverable rebuttal promises are cheap accepts; borderline papers with defensive rebuttals are cheap rejects.
Calibrating expectations by decision class
| Outcome signal | What it usually means | Your move |
|---|---|---|
| Conditional accept, short condition list | Committee wants the paper; conditions are the contract | Deliver precisely; see uist-camera-ready |
| Reject, reviews ask for more evaluation | Evidence gap — repairable in one cycle | Repair plan from the review bins |
| Reject, reviews dispute novelty | Related-work or delta failure | Fix positioning before anything else |
| Reject, reviews question the contribution's home | Venue mismatch | Re-run uist-topic-selection honestly |
After rejection: the three exits
- Adjunct fallback, same conference. A rejected paper's system is often a strong Demo or Poster submission to the same edition — the adjunct deadlines fall after paper notification. This keeps the work visible, collects hallway review, and stakes a dated adjunct publication. Remember the 2026 rule: one adjunct track per work, not both.
- Repair and resubmit to UIST next year: use the review bins as the work plan; evidence-gap rejections are the most repairable class.
- Re-route when reviews reveal a venue mismatch — "the contribution is the
study, not the system" is a CHI/CSCW signal, and
uist-topic-selectionshould be re-run rather than argued with.
Confidentiality and conduct
Reviews, reviewer identities, and PC discussion are confidential; do not quote reviews publicly or attempt reviewer identification. Contact the program chairs (program2026@uist.org for the 2026 cycle) only for process failures — a genuinely missing review, a compromised anonymity situation — not to relitigate scores.
Output format
[Phase] under review / rebuttal / decided-conditional / decided-reject
[Review bins] factual <n> · evidence <n> · novelty <n> · taste <n>
[Leverage estimate] rebuttal-winnable points and why
[If rejected] adjunct fallback / repair-resubmit / re-route — with the deciding review line
[Process flags] <anything warranting a chairs contact>
Version History
- 9f86f09 Current 2026-07-19 17:51
