Hello everyone,
we are in the phase of setting up and testing the SCRUM environment for our development projects. We'd like to make sure we're on the right track with the community's experience, and we'd really appreciate it if you could share your experiences and perhaps alert us to any issues or limitations in this part of the process that I'll describe below.
When designing this part of the process, we had the following goals:
- ensure that the Definition of Done is clearly visible to all members of the development team, but not in the form of notes or a list, but that the Definition of Done is converted into concrete tasks (sub-tasks),
- ensure that the DoD is set before starting work on a particular Story,
- automate the creation of sub-tasks according to DoD,
- get the possibility to easily filter tasks depending on the type of work (Development, QA, Documentation, ...), so that we can create separate panels, etc.
Considering the above, we are currently using (testing) this process:
1. At the Story level, we have defined a custom field "Definition of done (DoD)" by which the Product Owner defines what all actions need to be done in order to complete the Story. For example, the values can be:
- Code developed,
- Code reviewed,
- Unit tests complete and passing,
- Automated acceptance tests complete and passing,
- User documentation updated,
- Ops documentation updated,
- etc.
Some of the values are marked "true" by default.
2. When changing the status of the Story (currently to the status Ready to start), the automation that creates subtasks for the Story is triggered. The sub-tasks that are created have an issue type depending on the type of work to be done, so for example we can have sub-tasks of the following types:
- Dev Sub-task (for development, code revision, etc.),
- QA Sub-Task (for testing),
- Docs Sub-Task (for updating documentation).
Therefore, each member of the development team works on the Story exclusively through a sub-task. In this way, we have a clear picture of which part of the work is completed (or not).
3. We plan to add automation of status changes on Stories and sub-tasks that are dependent.
The dilemmas we have at the moment are the following:
1. Story points are of course estimated at Story level. But the Story itself remains assigned to the Product Owner, while each of the sub-tasks is assigned to someone else. Does this affect Sprint itself and in what way? Workload by assignee shows only people assigned to Stories, so when planning a sprint, we won't see the workload of other team members.
2. According to your experiences, is the direction described generally good or do you have suggestions for a better solution? What we want to keep is that each member of the development team has their own task that is part of the Definition of Done, for our three main areas of work: development, QA, documentation.
