Should we create a task for testing each testable development deliverable for should we test a development task before marking it as complete?
You should be testing a task before marking it as complete, otherwise it may become difficult to assess the state of a particular development task.
Additionally, it creates and artificial disconnected between your development tasks and the information they contain (such as user story, developer, any clarifications or discussions held on that task), and your test task (which may have it's own clarifications and discussions).
Testing is very much a two way process, so the tester should have knowledge of decisions the developer has made, and likewise the developer should see what the testers plan for testing is.
Our testers and developers discuss a task during planning, and the tester or developer will write some short QA Notes into the task upfront. Again keeping the testing activity and the development activity in the same place increases visibility across the two disciplines.
How do we mark it on the board in the latter scenario?
How we manage it, is to capture untested tasks and tested tasks as different states the ticket are in. We use jira, and have created a workflow with the following states:
ToDo -> InProgress -> PullRequested -> Merged -> Resolved
- ToDo: in the sprint backlog, ready to be worked on.
- InProgress: Developer actively working on task
- PullRequested: PullRequest open in git for team review
- Merged: Merged into feature branch and deployed to test environment. It's at this point that a tester will pick it up and test.
- Resolved: Tester has tested and accepted this task.
If a tester finds an issue with the task, he/she will put the tasks back into the ToDo state, and notify the developer of issues found. Keeping the testing activity captured in the dev tasks like this allows us to track the number of these 'reopens', with high levels of reopens being a sign of poor quality reaching the test team - an indicator of a problems.
Tim
manpreet
Best Answer
2 years ago
I was wondering what does the standard recommend as best practice in this scenario:
When creating tasks for a user story we usually have development/design/documentation/testing tasks. Should we create a task for testing each testable development deliverable for should we test a development task before marking it as complete?
Do we create separate testing tasks (which of course can not be started before the corresponding dev task is complete and will cause a certain task dependency and priority to develop) or do we simply consider a task complete once it's tested? How do we mark it on the board in the latter scenario?