percom-topic-selection
GitHub用于判断项目是否适合IEEE PerCom，或应路由至ACM UbiComp、MobiCom等会议。通过人类中心性测试和贡献类型区分PerCom与网络/传感器系统类 venue，并对比IEEE会议与IMWUT期刊的发表模式差异。
触发场景
安装
npx skills add brycewang-stanford/Awesome-Journal-Skills --skill percom-topic-selection -g -y
SKILL.md
Frontmatter
{
"name": "percom-topic-selection",
"description": "Use when deciding whether a pervasive-computing project belongs at IEEE PerCom or should be routed to ACM UbiComp\/IMWUT, MobiCom, MobiSys, SenSys, or IPSN, and when distinguishing PerCom's human-centric ubicomp focus from networked-sensor-systems venues by contribution shape, the human-centricity test, and the IEEE-conference vs. IMWUT-journal model."
}
PerCom Topic Selection
Decide the venue before drafting. PerCom — the IEEE International Conference on Pervasive Computing and Communications — is IEEE's flagship for human-centric ubiquitous computing: activity recognition, context-awareness, wearable and mobile sensing, and smart environments. Its papers are IEEE Xplore proceedings read for a pervasive-computing contribution grounded in real human use. A technically strong paper whose real lesson is about network protocols, low-power hardware, or a pure ML model is respected and then rejected as out of scope.
The routing question that matters most
PerCom overlaps most with ACM UbiComp/IMWUT in topic — both cover HAR, wearables, and smart spaces — so the decisive question is rarely "is this ubicomp?" but "which model and calendar fits this paper now?" UbiComp presents papers published in the IMWUT journal on rolling quarterly deadlines with a revise cycle; PerCom is an IEEE conference with a single annual deadline, a bounded rebuttal, and an in-person-presentation requirement. A strong ubicomp paper is publishable at either — choose by the publication model you want and the nearer honest deadline.
Neighbor-venue routing table
| Signal in your project | Better home | Why |
|---|---|---|
| Human-centric ubicomp (HAR, context, wearables, smart space), ready now, PerCom deadline nearer | IEEE PerCom | IEEE conference flagship; annual deadline + rebuttal |
| Same topic, but you want a journal publication and a rolling deadline | ACM UbiComp / IMWUT | Journal model, quarterly submissions, revise cycle |
| Core is wireless/cellular networking or the mobile network itself | MobiCom / MobiSys | Mobile-networking and mobile-systems centers |
| Core is a low-power networked-sensor system or sensor-platform design | SenSys | Networked sensor systems, not human-centric ubicomp |
| Core is information processing in sensor networks / signal-level sensing | IPSN | Sensor-network information processing |
| Study is a deep dataset/measurement better as an archival journal article | IMWUT / IEEE journals | Journal length and revise cycle |
Contribution shapes PerCom rewards
- Human activity / context recognition — a recognizer or model for human activity, context, or behavior from mobile/wearable/ambient sensors, evaluated cross-subject on real people (the EMGSense lineage).
- Pervasive system / infrastructure — a deployable system for sensing, localization, or smart environments whose insight outlives the specific hardware (the LANDMARC lineage).
- Sensing tooling / datasets — a generator, framework, or dataset that unblocks how the community evaluates (the SmartSPEC lineage).
- Sensing consequence / privacy — surfacing what ubiquitous sensing makes possible on real devices, for good or ill (the DeepMag side-channel lineage).
- Context modeling / device-to-device — treating context as a first-class systems concern with real deployment tradeoffs (the CHITCHAT lineage).
The human-centricity and model-swap tests
Two quick tests sharpen a borderline verdict:
- Human-centricity test: is a person — their activity, context, behavior, or environment — the center of the contribution? If the paper is really about the network stack or the sensor hardware with no human at the center, route to MobiCom/MobiSys or SenSys/IPSN.
- Model-swap test: if the paper leans on a deep model, ask whether the pervasive-computing lesson survives swapping the model for another. If not, the model is the contribution and an ML venue fits better; PerCom rewards the ubicomp insight, not the architecture.
Evidence maturity, without the ladder cliché
Fit is necessary but not sufficient: the same idea sits at different doors depending on evidence maturity. A recognizer evaluated only within-subject or in-lab needs cross-subject, free-living data before the research track; a system shown only on your own testbed needs a realistic deployment; a study too deep and long for 9 pages may belong in IMWUT first. Submitting a step early earns a polite "promising, but the evaluation is within-subject" and costs a full cycle, because PerCom runs on an annual rhythm.
Cheap reconnaissance before committing
[Scope] scan the last two PerCom programs (dblp, percom.org) for your subarea
-> 3+ recent papers = a reviewer pool exists; 0 = opening or mismatch
[Citations] is your bibliography majority ubicomp/mobile/sensing venues (PerCom/UbiComp/IMWUT/
MobiCom/SenSys)? -> majority non-ubicomp => reviewers read you as a visitor; naturalize first
[Calendar] compare the next PerCom deadline (annual, ~Sep) with IMWUT's rolling quarterly cutoffs
and MobiCom/SenSys dates -> route to the nearest honest fit, not a year's wait
Decision procedure
[Audience] who acts differently if the claim holds? -> ubicomp researchers/practitioners/users?
[Claim type] recognition / system / tooling-dataset / sensing-consequence / context-modeling
[Human-centricity] is a person at the center? -> yes: PerCom-family; no: MobiCom/SenSys/IPSN
[PerCom vs IMWUT] both fit? -> conference+rebuttal+annual (PerCom) vs journal+rolling (IMWUT)
[Verdict] PerCom research track / neighbor venue / journal-first, with a one-line reason
Run this before the writing skills; a wrong venue decision wastes every later step. When the
verdict is PerCom, continue with percom-workflow for the calendar and percom-writing-style for
the paper shape.
版本历史
- 9f86f09 当前 2026-07-19 17:19
