Studio culture & creative process / DIJION
From It does not work to a reproducible mobile bug report
A bug-report plan separating initial conditions, steps and expectations, with interruption, device and data context.
Save does not work may describe an important problem without letting another person reproduce it. In an illustrative notes app, the failure might occur only in an older draft edited offline. A useful report turns the complaint into investigable conditions, behavior and consequences without requiring the person reporting it to diagnose the implementation.
Record initial conditions before the steps
Account type, record state, connectivity and build can change the outcome. Open the app may be insufficient: was it already running or starting a new process? Describe necessary data through a synthetic example. Passwords, session credentials and real private documents should not become prerequisites for reproducing an issue.
Keep expected and actual results distinct
Scroll the table horizontally to see all columns.
| Field | Useful wording | Wording to avoid |
|---|---|---|
| Step | Change the title offline and choose Save | Try things for a while |
| Expected | Draft is retained and state is explained | It should work properly |
| Actual | Text disappeared without an error | The database must be broken |
Match evidence to the question
A short recording can show sequence while a screenshot captures a visible result. Remove private notifications or information before sharing. Limit logs to the relevant interval and safe fields. Record repeat frequency without marking untested devices as affected; keep the verified environment distinct from the suspected scope.
Do not present the symptom as the cause
Record which condition changes when simplifying the reproduction. A failure disappearing with different connectivity does not prove a network cause; timing may also have changed. Justify whether multiple symptoms belong in one investigation. Let the title describe the user-visible impact and keep possible implementation explanations in a separate research note.
Have another person follow the report
- Someone without prior context should establish the same initial state.
- Steps should specify both sequence and action.
- Expected result, actual result and assistance needs should be separate.
- Verify the fix against the same scenario and a related flow.
The outcome is not a longer ticket but one requiring fewer assumptions. Even when reproduction fails, preserve the observed evidence and mark the missing context explicitly.