Use this as the final audit before a paper enters an S&P cycle on HotCRP. S&P's desk-reject rules are unusually mechanical — page counts, ORCID matching, and registration completeness are enforced without review — so most preventable losses happen in the upload week, not in review.
An S&P cycle has a registration deadline and a paper deadline about a week apart, and the registration is binding: title, full abstract, complete author list, ORCIDs, and conflicts are frozen when registration closes (cycle1.sp2027.ieee-security.org/deadlines, checked 2026-07-08). Audit registration as its own deliverable:
ieeesp-review-process for how the REC uses it).SoK: title prefix and the SoK checkbox — one
without the other misroutes reviewing.| Rule (2027 CFP, checked 2026-07-08) | Enforcement |
|---|---|
| ≤ 13 pages of body text | Over-length ⇒ rejected without review |
| ≤ 5 further pages, references + appendices only | Same |
| Hard ceiling 18 pages total | Same |
| Everything past page 13 clearly marked as appendix | Reviewer-visible violation |
| IEEE "compsoc" conference-proceedings LaTeX template | Tampering ⇒ desk-level flag |
| Anonymized submission | Identity leak ⇒ desk reject |
Two S&P-specific implications. Reviewers are not required to read appendices, so a proof, an ablation, or the adaptive-attack evaluation that the acceptance case depends on must live in the first 13 pages. And the compsoc template's two-column layout eats wide tables and long listings — compress artifacts early, not on deadline night.
Security submissions leak identity in ways generic checklists miss:
/home/<user>/...) all deanonymize.# Mechanical pass before upload
pdfinfo paper.pdf | grep -iE 'author|creator' # metadata
pdftotext paper.pdf - | grep -inE 'our (prior|previous) work|we reported|CVE-20'
grep -rniE '(\.edu|\.ac\.|corp)/|/home/[a-z]' figures/ scripts/ || true
| Trigger | Repairable after deadline? |
|---|---|
| Missing/mismatched ORCID for any author | No — desk reject |
| Body text past page 13 or total > 18 | No — rejected without review |
| Author list change after registration | No — frozen |
| Identity leak in PDF or artifact | No |
| Weak Ethics Considerations entry | No new entry; damage lands in review/REC |
| Thin adaptive-attack evaluation | Review-stage damage; fix before, not after |
[S&P upload readiness] Ready / Needs fixes / Not ready
[Registration] authors+ORCIDs ✓/✗ · abstract ✓/✗ · conflicts ✓/✗ · ethics field ✓/✗
[Format] body ≤13 ✓/✗ · total ≤18 ✓/✗ · compsoc ✓/✗ · appendix marked ✓/✗
[Anonymity] pdf metadata / disclosure trail / artifact origins: <findings>
[Highest desk-reject risk] <one item>
[Fix order] <ordered list before the registration deadline>