跳到主要内容
知仓学习社ZHICANG

mobicom-artifact-evaluation

Use when packaging a MobiCom artifact for the evaluation committee — choosing among the three ACM badges as a calibration of what you can prove, bui…

不碰外部(只输出文字)无严重或高危命中brycewang-stanford/Awesome-Journal-Skills

它会碰到什么

扫了多少1 个文本文件,5 KB
它会碰到什么不碰外部(只输出文字)
命中总数0 处
命中统计严重 0 · 高 0 · 中 0 · 低 0

这一栏是扫描器报的事实,不是结论。命中多不等于有毒(安全工具、规则库、示例脚本本来就会包含危险写法),命中少也不等于干净。它和你手上的凭据、文件、网络有什么关系,需要你自己看。

技能内容

MobiCom Artifact Evaluation

MobiCom accepted papers may earn the three ACM badgesArtifacts Available,

Artifacts Evaluated — Functional, and Results Reproduced — from a separate

artifact-evaluation committee (AEC). The badges are independent: you may pursue one, some, or

all. The engineering problem specific to MobiCom is that the evaluation depends on a **radio

testbed the AEC does not have**, so the packaging must let an evaluator get as far as

possible without one.

Choose badges as a claim of what you can prove

| Badge | What it asserts | What the AEC needs |

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

| Artifacts Available | the artifact is publicly and permanently retrievable | a stable public archive (DOI-backed) |

| Artifacts Evaluated — Functional | the artifact runs and does what the paper describes | a working build + smoke run on reachable hardware |

| Results Reproduced | the paper's key results regenerate | a path from artifact to the headline numbers |

Pick badges honestly. Available is almost always worth claiming and is the lowest-friction.

Functional requires a path that runs without your specific radios. Results Reproduced is

hardest for a wireless paper because the channel is not portable — claim it only where a

trace-replay or downscaled path can regenerate the numbers (mobicom-reproducibility). The

exact 2027 badge deadlines and process are 待核实 — read the current artifact call.

The hardware-optional path is the core deliverable

Most AEC members will not have your SDRs, tags, or testbed. Design the package so a reviewer

without hardware still reaches Functional:

artifact/
  README            # claims, badge targets, hardware needed vs optional, time budget
  env/              # container or pinned deps; no radio required to build
  smoke/            # runs in ~minutes on a laptop, no radio; proves the pipeline
  replay/           # recorded captures + decode/analysis -> regenerates key figures
  hardware/         # optional: scripts + exact radio/firmware to reproduce on a testbed
  data/             # traces that can legally ship (or a loader + instructions)
  expected/         # expected outputs AND expected variance ranges
  • The smoke run is the functionality proof: a short, laptop-only command that exercises

the pipeline end-to-end on bundled data and prints a recognizable result.

  • The replay path feeds recorded captures through the same processing that produced the

paper's figures, so results regenerate without the radio.

  • The hardware path is documented for the rare evaluator (or future reader) with a

matching testbed, with exact radio model, firmware, and channel setup.

Make availability real

  • Deposit in a permanent, DOI-backed archive, not a personal URL that rots.
  • Confirm what can legally ship: firmware and proprietary radio stacks often cannot;

user-identifiable captures must be stripped or aggregated; site data may be bound by an

agreement (mobicom-reproducibility).

  • If part of the artifact cannot be released, say so explicitly and ship the largest

functional subset rather than a broken whole.

Document variance honestly

An over-the-air result differs run to run. Ship expected-variance files alongside

expected outputs so a rerun that lands within a stated margin reads as a reproduction, not a

failure. An AEC member who sees a number move and has no variance guidance will log it as a

non-reproduction.

Pre-submission checklist

  • [ ] Badge targets chosen to match what the package can actually prove.
  • [ ] Smoke run works on a clean laptop with no radio, in a stated time budget.
  • [ ] Replay path regenerates the key figures from bundled captures.
  • [ ] Hardware path documents exact radio/firmware/channel for testbed reproduction.
  • [ ] Available artifact is in a permanent DOI-backed archive.
  • [ ] Legal/ethical shipping decision made for firmware, traces, and site data.
  • [ ] Expected outputs and expected variance both provided.

Output format

[Badges] targeted vs justified by the package
[Smoke] laptop-only run present + time budget
[Replay] key figures regenerate without hardware? y/n
[Hardware path] radio/firmware/channel documented? y/n
[Availability] archive + DOI + legal-shipping decision
[Gaps] what blocks each targeted badge

想直接用这个技能?

本站把开放许可(MIT / Apache 等)的技能按仓库打包整理到网盘,点一下转存到你自己的网盘,不用一个个从 GitHub 拉。许可未声明的技能只给原始仓库链接,不打包。