| name | uist-submission |
| description | Use when performing the final pre-upload audit of a UIST paper in PCS — the abstract-then-paper deadline pair, 10-page/5-page two-column limits with desk-reject enforcement, anonymous acmart options, video-figure and supplementary uploads, the prior-work overlap rule, and deadline-week sequencing. |
UIST Submission
This is the audit to run in the final week before the Papers deadline. Numbers below
are the verified 2026 cycle (checked 2026-07-08 via search renderings of
uist.acm.org/2026/, direct fetch blocked): abstracts March 24, papers March 31,
both 11:59 pm AoE. The 2026 deadline has passed — apply this skill to a 2027
submission only after re-reading that year's Author Guide, and expect the numbers to
move.
What PCS collects, and when
UIST runs on PCS (Precision Conference Solutions): in the submission form, choose
society "SIGCHI", conference "UIST 2026", track "UIST 2026 Papers". Two deadlines
structure the week:
- Abstract deadline (one week out). Title, abstract, author list, and metadata
must exist in PCS. This gates reviewer assignment — a placeholder abstract that
drifts from the final paper lands you with mismatched reviewers, the least
fixable submission error.
- Paper deadline. The PDF, the video figure, and all supplementary files.
Everything reviewed arrives now; there is no later supplement slot.
Format gate
| Check | 2026 rule | Enforcement |
|---|
| Template | Two-column acmart, \documentclass[sigconf,review,anonymous]{acmart} | Manual + visual |
| Body length | 10 pages (standard) or 5 pages (short), excluding references and appendices | Desk reject if exceeded |
| Appendices | Inside the same PDF, after references; body must stand alone | Review-stage damage |
| Video figure | Optional, strongly encouraged: ≤ 3 min, 1080p/4K, MP4 H.264 | Reviewer-side (see uist-supplementary) |
| Anonymity | Paper, video, and supplement all purged of identity | Desk reject |
A short paper is a distinct category (smaller complete contribution), not a truncated
standard paper — do not switch categories on deadline night to dodge the page count.
Anonymity, the UIST version
The standard is "reasonable effort," with two venue-specific twists:
- Third-person self-citation with full references. Do not redact your own
citations; write about your prior work as if someone else built it.
- Overlap attachment. If a prior submission of yours overlaps significantly,
upload an anonymized copy of it as supplementary material.
Sweep everything that ships: PDF text, PDF metadata, acknowledgements and grant
numbers, repository and project-page URLs, video frames (lab posters, faces,
institutional wallpaper), video audio (narrator self-identification), file names,
and archive contents.
pdfinfo paper.pdf | grep -Ei 'author|creator|producer'
pdftotext paper.pdf - | grep -nEi 'github|gitlab|osf\.io|acknowledg|grant' | head
ffprobe -v error -show_entries format=duration:stream=width,height,codec_name video.mp4
unzip -l supplement.zip | grep -Ei '\.git|DS_Store|/Users/|/home/' | head
Deadline-week order of operations
- T-7d: abstract deadline — final title/abstract in PCS, authors and conflicts
complete, subject metadata chosen for reviewer routing.
- T-5d: body frozen at the page limit; appendix absorbs the overflow.
- T-4d: video locked (edits, captions, anonymization) and compressed to spec;
verify PCS accepts the file size early, since caps were not published (待核实).
- T-2d: full anonymity sweep on the merged PDF, video, and archive; run the
mechanical checks above on the exact files to be uploaded.
- T-1d: upload everything, then re-download from PCS and open cold — the
re-downloaded copy is what reviewers see. Confirm the PCS abstract matches the
PDF abstract word for word.
- T-0: buffer only. AoE means UTC-12; convert once, write it on the wall.
PCS form fields worth drafting early
The form is read before the PDF and configured per track; draft its contents days
ahead rather than composing in the upload session:
- Title and abstract — identical to the PDF, since the abstract drives
reviewer matching from the abstract deadline onward.
- Author list and conflicts — complete at abstract time; adding authors later
in the process is at minimum awkward and possibly restricted (待核实 per cycle).
- Subject/topic descriptors — choose for the reviewer pool you want; a
fabrication paper tagged only "AI" gets AI reviewers.
- Prior-submission declarations — where the form asks about overlapping or
previously reviewed material, answer literally; this is also where the
anonymized overlap attachment gets referenced.
- Track category — standard vs short is a claim about contribution scale;
reviewers hold you to the category you picked.
Last-day failure modes, ranked by frequency
| Failure | Prevention |
|---|
| Video upload stalls or exceeds an unpublished cap | Upload media a full day early; keep a lower-bitrate export ready |
| Final LaTeX rebuild shifts the paper past the limit | Freeze the body two days out; late edits go to the appendix |
| PDF metadata carries an author name from a home template | Mechanical sweep on the exact export, every export |
| PCS abstract diverges from a last-minute PDF abstract edit | One person owns both texts; diff them at T-1d |
| A co-author's "tiny fix" reintroduces an acknowledgements block | Lock the repo after the anonymity sweep |
Dual-track and dual-venue rules
- Papers under review elsewhere may not be under review at UIST at the same time.
- Within UIST 2026, the same work sent to both Demos and Posters is treated as a
demo submission and rejected from Posters — pick one adjunct lane deliberately.
- A prior adjunct publication (demo/poster abstract) of the same system does not
anonymize away: cite it in the third person and state the delta.
Output format
[Submission status] ready / blocked / at risk
[Format gate] pages <n>/<limit> · template ok? · video spec ok?
[Anonymity sweep] clean / leaks: <file: location>
[PCS state] abstract locked · uploads verified by re-download?
[Remaining actions] <ordered, with owner and hour>
