---
name: mobisys-workflow
description: Use when planning a MobiSys project timeline backward from the single early-December paper deadline through registration, two-round review with an early-reject cut, the rebuttal, notification, artifact evaluation, camera-ready, and the June conference — with device-experiment lead time built in and the one-deadline-per-year risk made explicit.
---

# MobiSys Workflow

Use this as the project-management skill for a MobiSys submission. Replace every date with the
current official timetable and work backward from the one hard December cutoff on the
`mobisys<yy>.hotcrp.com/deadlines` page.

MobiSys is a conference, not a journal: no standing editor-in-chief, no article-processing
charge. The rotating leadership is the per-edition General Chairs and Program Chairs (2026
General Chairs: Fahim Kawsar and Cecilia Mascolo; Program Chairs 待核实). Committees rotate
yearly — re-check the current organizing-committee page rather than carrying a name forward.

## The one-deadline calendar

Unlike MobiCom and NSDI, which run two deadlines a year, **MobiSys runs a single paper
deadline in early December feeding the following June conference**. There is no second gate
this cycle to catch a slip. That single fact governs the whole plan: the device experiments
must be finished, not started, before December.

## Milestones

- Venue fit: confirm the contribution is a mobile/embedded *system* whose device behavior is
  the result (`mobisys-topic-selection`).
- Evidence lock: freeze the system design and finish the on-device measurement campaign —
  energy, latency, memory, thermal, and any deployment/user study.
- Paper registration: register title, abstract, authors, and conflicts on HotCRP before the
  registration cutoff that precedes the paper upload.
- Paper deadline: upload the anonymous 12-page double-column PDF and required HotCRP fields.
- Round-1 outcome: early-reject notifications go out; survivors proceed to round 2.
- Round-2 reviews + rebuttal: read reviews, then draft a scope-limited rebuttal
  (`mobisys-author-response`).
- Notification: archive reviews, rebuttal, and decision.
- Acceptance: prepare the ACM camera-ready, artifact-evaluation submission, registration, and
  the June in-person talk and demo.

## Backward plan from the December deadline

| Weeks out (heuristic) | Mobile-systems milestone |
|---|---|
| 12+ | System built; measurement harness and power instrument working end to end |
| 8 | On-device runs complete on all target devices; energy/latency logged with provenance |
| 6 | Deployment or user study done if the claim is a service |
| 4 | Full draft sitting in the 12-page ACM double-column template |
| 3 | Internal mock review by a systems reader; break-condition experiment added |
| 2 | Anonymity sweep including device/trace giveaways; artifact package smoke-tested |
| 1 | Numbers regenerated from logs so PDF and artifact cannot diverge |
| 0 | Registration then paper on HotCRP before the UTC cutoff |

These offsets are planning heuristics only — anchor each to the current official timetable,
never to a previous cycle's calendar.

## Failure modes by stage

- Starting device measurement in November forces single-device, single-run results that read
  as under-evaluated — the classic MobiSys evidence reject.
- Leaving format conversion to the final week surfaces overflow late; the 12-page
  double-column budget is tighter than a working draft once figures are placed.
- Skipping the mock review forfeits the only chance to hear "your energy boundary is
  undefined" from a colleague instead of a reviewer.
- Building the artifact after the paper deadline forfeits the ACM badges, which need a working
  script-first package the AEC can run on its own machines (`mobisys-artifact-evaluation`).

## Coordination notes

- Assign one named owner for the on-device measurement provenance and another for the
  anonymity sweep; shared ownership is how both slip.
- Archive the exact submitted PDF, supplement, and artifact commit, since rebuttal replies
  must quote them precisely and camera-ready must match the reviewed version.
- Because a miss costs a full year, decide the go/no-go on evidence completeness by the
  four-week mark, not on deadline night.

## Output format

```text
[Current stage] idea / building / measuring / writing / registration / submitted / rebuttal / accepted
[Next official deadline] <date and source, or unknown>
[Critical path] <three tasks that determine readiness>
[Risk register] <device evidence / page budget / anonymity / artifact / one-deadline slip>
[Owner map] <task -> person or role>
```