Use this before submission and again before camera-ready. IMC treats availability and reproducibility as scored dimensions, not courtesies: the submission carries an artifact-availability declaration, accepted papers are shepherded to deliver it, and IMC runs a dedicated Replicability Track. The goal is that a competent reader could rebuild your analysis from the released data — and re-run your method to gather comparable new data — and reach your conclusions.
Measurement differs from lab science: you cannot re-collect the same data, because the network changes between runs. So reproducibility at IMC splits in two:
Say which you provide, and never present method reproducibility as if it reproduced your exact numbers.
| Claim in the paper | Weak answer | IMC-ready answer |
|---|---|---|
| "We scan N hosts" | "Data available on request" | Released dataset (or privacy-safe derivative) + scan metadata + scripts |
| "Measured from many vantage points" | Vantage points unnamed | Documented vantage-point table: locations, ASes, probe types, dates |
| "We observe behavior X over time" | Single-snapshot data | Time-stamped longitudinal data + the analysis window stated |
| "Our tool detects Y" | "Code will be released" | Released, runnable tooling with a README and a small example |
| "We used user/traffic data" | Nothing (privacy cited vaguely) | Aggregated/anonymized release + documented privacy method + ethics link |
[Vantage points] record every location/AS/probe-type, and quantify coverage bias
[Timing] measurement dates, durations, cadence; state the analysis window
[Targets] target/seed lists with capture dates; how they were sourced (and their stability)
[Tools] exact tool versions and configs; rate limits and opt-out/blocklist handling
[Sampling] inclusion/exclusion criteria and resulting sample sizes; how you handle churn
[Privacy] the anonymization/aggregation applied before release, matching the Ethics section
Aim turnkey for the analysis of any releasable dataset; when data is sensitive, provide a privacy-safe derivative plus method-only reproducibility, and say so plainly.
IMC runs a dedicated Replicability Track for work that reproduces or replicates prior measurement results, entered via an Expression of Interest screened by a small committee, then a full submission judged like the main track (priority to replicability over reproducibility). If your contribution is re-measuring a prior study on today's Internet, this is the track — design for a fair, documented re-run and an honest account of what changed and why.
imc-artifact-evaluation).[Claim inventory] <claim -> evidence location>
[Availability] full / partial / none, stated honestly (+ reason)
[Provenance gaps] <vantage points / timing / targets / tool versions / privacy>
[Reproducibility level] turnkey-analysis / scripted / method-only, stated honestly
[Paper fixes] <must appear in the PDF>
[Release fixes] <additions before upload / camera-ready>