agentsclimarketplace

Hpca review process

Skill brycewang-stanford/Awesome-Journal-Skills/HPCA-Skills/skills/hpca-review-process

Use when reasoning about the HPCA review pipeline: the double-blind process, what a single combined rebuttal/revision window means, champion dynamics at the program-committee meeting, how to read each outcome, and how confidentiality and the preprint policy shape expectations.From its SKILL.md

Install
npx -y skills add brycewang-stanford/Awesome-Journal-Skills --skill hpca-review-process

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

SKILL.md

3.7 KB, 788 tokens by cl100k_base, as published. Nobody here has run it

HPCA Review Process

Use this to set expectations about how HPCA reviews a paper and decides its fate. Understanding the pipeline turns anxiety into a plan: you can predict what the rebuttal/revision window is for and what each outcome signals.

The pipeline

HPCA runs a double-blind review. After the late-July gate, papers are assigned to program-committee reviewers, reviews are written through the autumn, and — in recent cycles — authors get a single combined rebuttal/revision window: one period in which you both respond to reviews and may revise the paper, rather than two separate review rounds. Decisions follow a program-committee meeting; HPCA 2027's reported notification is November 6, 2026 (待核实 the exact rebuttal window).

What the single window means

Unlike venues with two review rounds, HPCA's combined window compresses "argue" and "change the paper" into one effort. Plan for both at once:

  • Reviewers see your response and a revised paper, so a claim you defend in prose should also be reflected in the revision.
  • The window is short; the objection ledger and staging shelf (see hpca-author-response and hpca-supplementary) must be ready before it opens.

Champion dynamics at the PC meeting

Architecture decisions are made by people in a room (or a call). A paper needs a champion — a reviewer willing to argue for it — and its fate often turns on whether the champion can answer the strongest objection. Write the rebuttal to arm your champion with the room, not only to satisfy the skeptic.

OutcomeLikely meaningYour move
AcceptThe champion held; objections were addressableCamera-ready + artifact evaluation
Reject with strong reviewsA fixable but unaddressed methodology or novelty gapFix, then retarget MICRO/ISCA/ASPLOS
Reject with weak/split reviewsNo champion emerged, or scope mismatchReframe the contribution, re-route
Borderline / conditionalThe room was dividedRead every review for the deciding objection

Read reviews for the deciding objection

Not all objections weigh equally. Identify the one that decides the paper — usually a methodology or novelty concern the champion must overcome — and prioritize it over easy-to-fix surface comments. A rebuttal that answers three minor points and dodges the deciding one loses.

Confidentiality and preprints

Reviews and PC discussion are confidential; do not solicit or reveal reviewer identities, and do not de-anonymize yourself mid-cycle. Recent HPCA guidance asks PC members not to penalize arXiv/CAL preprints, so a preprint is allowed — but posting one that reveals authorship does not change the double-blind obligation on the submission itself.

Expectation-setting pass

[Stage] assigned / reviews out / window / PC decision / notified
[Window plan] response + revision both staged? (Y/N)
[Deciding objection] <the one that decides the paper>
[Champion arming] does the rebuttal give the champion the room? (Y/N)
[Outcome branches] accept / reject-strong / reject-split -> next action

Output format

[Pipeline position] <where the paper is>
[Rebuttal/revision window] <dates from the current CFP, or unknown>
[Priority objection] <one>
[Outcome-conditional plan] <accept / reject branches>

Reopen the current CFP for the exact review calendar and window dates — the pipeline timing is per-edition and was unconfirmed for 2027 at pack time.

What ships with it

Read from the repository

Just SKILL.md. No reference files, no scripts.

Keep looking

Skills are one crate of 325,949. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.