| name | isca-camera-ready |
| description | Use when converting an accepted ISCA paper into its published form — de-anonymizing authors, acknowledgments, and artifact links; satisfying the edition's publisher workflow under the alternating ACM/IEEE arrangement; integrating shepherd requirements and badges; and preparing the June talk and attendance logistics. |
ISCA Camera-Ready
Acceptance starts a second, administrative project with real failure modes:
papers have appeared with anonymization stubs intact, dead artifact links, and
acknowledgments naming the wrong grant. The 2026 edition maintained a dedicated
camera-ready guidelines page (iscaconf.org/isca2026/submit/cameraready.php —
its detailed contents were not directly fetchable at verification time, so the
specific instructions below marked practice are community-standard rather than
quoted rules; the page itself is the authority — 待核实 per edition).
The publisher variable unique to ISCA
ISCA's ACM SIGARCH + IEEE-CS TCCA co-sponsorship shows up concretely at
camera-ready: the publishing arrangement varies by edition (the 52nd
edition's proceedings were ACM-published; the 2026 edition carried IEEE
branding and pointed LaTeX users at an IEEE template). Everything downstream —
rights forms, copyright blocks, final template, DOI issuer, submission portal
for finals — follows the edition's publisher. Rule one of ISCA camera-ready:
identify this edition's publisher workflow before touching the LaTeX, and
expect it to differ from the last ISCA you published at.
De-anonymization is a checklist, not a mood
Work from the submission's blinding actions and invert each one deliberately:
| Was blinded at submission | Camera-ready action |
|---|
| Author block absent | Full names, affiliations, emails — order confirmed by all coauthors in writing before upload |
| PDF metadata scrubbed | Regenerate with true metadata; verify with pdfinfo |
| Artifact links anonymized | Swap in canonical repo + archival DOI (isca-artifact-evaluation); click every link from the final PDF |
| Third-person self-citations | Recast where the prose reads absurdly; keep where fine |
| Acknowledgments omitted | Add funders with exact grant numbers, contributors, and any disclosure the publisher's policy requires |
| Neutralized cluster/trace names | Restore real names where licenses permit |
pdfinfo final.pdf | grep -i author
pdftotext final.pdf - | grep -ci 'anonym'
pdftotext final.pdf - | grep -nE 'blind review|under submission'
pdftotext final.pdf - | grep -oE 'https?://[^ )]+' | sort -u > links.txt
while read -r u; do curl -s -o /dev/null -w '%{http_code} %s\n' "$u" "$u"; \
done < links.txt | grep -v '^200'
Shepherds and required changes
If the acceptance carries a shepherd or a mandatory-revisions list, that list is
a condition, not a suggestion. Practice that keeps this smooth: send the
shepherd a change memo (item → what changed → where) with the revised draft,
early enough for one round of disagreement; where you believe a requested change
is wrong, negotiate in the memo — never silently skip an item and hope.
Content changes: repair, don't remodel
Camera-ready is for review-driven repairs and accuracy fixes. If numbers
changed (a bug found post-acceptance, a rerun under the fixed seed), correct
them and ensure abstract/intro claims still match the tables — publishing known-
stale numbers because "it was accepted with them" is how errata happen. Do not
add new contributions; the version of record should be the reviewed paper,
healed. Any camera-ready page allowance beyond the submission's 11 text pages is
edition-specific and was not verified — 待核实 before planning to grow.
Badges, statements, and the artifact block
Coordinate with the AE track (isca-artifact-evaluation): earned badges are
applied to the published paper under the ACM badging framework, and the
camera-ready typically carries an artifact-availability statement with the
archival DOI. Get the DOI before the camera-ready deadline by depositing the
evaluated snapshot early; retrofitting a DOI after the proceedings freeze is
somewhere between painful and impossible.
The June obligations
The 2026 conference ran June 27 - July 1 in Raleigh. Between camera-ready and
the conference, in rough order of deadline pressure:
- Registration — whether a specific author-registration category is
mandatory per paper was not verified (待核实, and commonly true at flagship
venues); resolve it the week of acceptance, with visa lead times in mind for
international authors.
- Talk — the community's norm is mechanism-forward: motivate with your
measured data, spend the middle on how the design works (the walk-through
from
isca-writing-style translates directly to slides), end on the honest
limits. Rehearse against the session's time slot, not a guess at it.
- Poster/lightning collateral if the edition uses them — 待核实.
- Publicity — once the proceedings version exists: update the arXiv record
if one exists, the lab page, and the repo README to cite the canonical DOI,
so the version of record accretes the citations.
Sequence for the six-or-so weeks
Week 0 Read THIS edition's camera-ready page top to bottom; identify
publisher, template, rights form, portal, true deadline (AoE?).
Week 1 Shepherd memo (if any) + content repairs; coauthor sign-off on
author order and affiliations.
Week 2 De-anonymization checklist; acknowledgments with grant numbers;
artifact DOI deposited.
Week 3 Publisher template pass; rights form executed by the authorized
author; metadata (title case, abstract, keywords) entered exactly.
Week 4 Link check + verification script above on the actual final PDF;
upload; download back and re-verify what the portal stored.
Buffer Publisher back-and-forth (format rejections are routine, not
emergencies) + registration + talk drafting.
Final gate before upload
Verified anchors (dates, sponsorship, AE arrangement, 2026 template flavor):
../../resources/official-source-map.md, checked 2026-07-08. The camera-ready
page contents, page allowance, rights workflow, and attendance rules for the
target edition are the controlling sources once posted.
