imc-review-process

IMC Review Process

Model the pipeline before interpreting any single review. IMC's process is built around **two

deadlines a year and a distinctive middle outcome — One-Shot-Revision** — that is neither a

plain reject nor an open-ended major revision. The most consequential mental shift for authors

arriving from a single-round conference is that a good-but-incomplete measurement paper has a

bounded path to acceptance across deadlines.

Process model

• Submission and review run on HotCRP with double-blind anonymity maintained through

reviewing.

• Review proceeds in several rounds. Papers with a **consensus for rejection are notified

early**, so authors can move the work to another venue without waiting for the full cycle.

• First decisions fall into Accept, One-Shot-Revision, or Reject. Reviewers weigh the

significance of the measurement finding, the credibility of vantage points and ground truth, the

soundness of methodology, honesty about limitations and confounds, the Ethics treatment, and

artifact availability.

• Accepted papers may be shepherded by PC members — notably to ensure the **artifacts promised

in the availability declaration are actually made available**.

The One-Shot-Revision, precisely

| Element | What the call specifies |

|---|---|

| Trigger | At least one reviewer champion advocating the work's relevance and timeliness |

| Scope | Up to 3 additions/changes — a bounded, concrete action list, not "revise broadly" |

| Resubmission | To the next deadline in the two-deadline cycle |

| Re-review | By the same reviewers, against the action points |

The strategic reading: a One-Shot-Revision is a contract. It is granted because someone

believes in the paper, and it is judged narrowly on whether the named action points were done.

Treat it as a near-accept with obligations, not a coin flip (see imc-author-response).

Reading a decision against the categories

| Decision | What it means | Author move |

|---|---|---|

| Accept | Finding and method hold; shepherding for availability | Deliver the promised artifact; do not reopen scope |

| One-Shot-Revision | A champion exists; a bounded set of gaps (a missing vantage point, a confound to address, an ethics clarification) | Execute exactly the named action points; resubmit next deadline to the same reviewers |

| Reject | Structural: unrepresentative vantage points, no ground truth, thin measurement, ethics failure | Reframe or reroute (SIGCOMM/NSDI/CoNEXT/PAM or a journal); do not lightly resubmit unchanged |

Design the submission so its weakest point is an addressable action item (an extra vantage

point, a longitudinal slice, a bounded confound) rather than a structural flaw (a study design you

cannot redo before the next deadline).

How IMC differs from its siblings

• vs. SIGCOMM / NSDI: those reward systems and protocol design with a supporting evaluation;

IMC rewards the measurement itself and enforces the Ethics gate and availability declaration

harder. Never assume a shared calendar or template.

• vs. a single-deadline conference: IMC's two deadlines plus One-Shot-Revision give a *within-

year* second chance with continuity of reviewers — closer to a lightweight revise-and-resubmit

than to "reject and try next year."

• vs. a journal: the revision is one shot and bounded (up to a few action points), not an

open iterative loop.

Who reads you

Expect measurement experts who will interrogate vantage-point representativeness, whether

ground truth is real, whether the finding is a snapshot artifact of a moving Internet, and whether

the active measurement was safe and ethical. Because IMC spans scanning, DNS, routing, web,

security, and social measurement, papers are matched to subarea reviewers — a vague methodology

gets caught, not skimmed.

Where author leverage actually exists

[Before submission]  topic tags + vantage-point framing -> reviewer pool      (largest lever)
[Initial reviews]    factual corrections, clarifying vantage-point coverage, ethics clarifications
[One-Shot-Revision]  the strongest lever: execute the bounded action points exactly, for the
                     same reviewers, and show each one done
[Shepherding]        deliver the promised artifact; this is about availability, not new science
[After reject]       no appeal; reroute to a sibling venue or a journal

Misreadings to avoid

• Treating One-Shot-Revision as a full rewrite window — you get a bounded action list and the

same reviewers; scope creep signals you missed the point.

• Treating One-Shot-Revision as guaranteed — the second read against the action points is

real.

• Ignoring the early-reject signal — a consensus-reject notice is a routing prompt, not an

invitation to argue.

• Underestimating shepherding — a promised dataset that never materializes is a real problem.
• Projecting last year's cadence — deadline dates and revision mechanics are per-edition.

Output format

[Process stage] pre-submission / awaiting reviews / one-shot-revision / shepherding / accepted
[Decision category] accept / one-shot-revision / reject, with the criterion driving it
[Champion?] is there a reviewer advocate? which review reads closest?
[Action points] <the bounded list to execute for the next deadline>
[Leverage plan] the next-stage action that can actually change the outcome
[Forbidden moves] identity leak / scope creep beyond the action points / undelivered artifact