Sprint 3 goal: The goal for the third sprint of Project Ernst was to start working on some of the biggest structural changes ever made to our platform. In other words, the largest structural obstacle should be partially or completely neutralized before the end of the sprint.
Not familiar with Project Ernst? Check out part 1to 3 here: Project Ernst, the first sprint and the second sprint. Project Ernst aims to create ownership of each team’s code base. The story is told as it happens.
Confronting our biggest structural obstacle meant that the different teams once again had different issues to address. But they all had one thing in common, all the teams had chosen areas that generate large amounts of support issues, inconsistency for users and that was time-consuming to deal with for the teams.
Launching a two-stage rocket
One of the teams chose to tackle a part of their domain that is handling sorting, which over time has become complex and with poor scalability. Because this part is crucial for the team’s domain, it was decided to attack the problem in a two-stage rocket, first abstracting the sort logic completely, to replace it completely at a later stage.
A new algorithm
A underlying reason why the team chose this area was that Peter Klaesson, during one of our “innovation days”, had developed a proof-of-concept for a new sorting algorithm. The team thought this area was a priority and they had a clear idea of the target image.
Pushing the limit
As the task was a big deal, the team and the product owner was initially prepared to continue the work even after the last sprint in the project was over. When this text is written, the work with abstracting the sort logic is still in progress.
Improving the data model
The other team addressed a similarly time-consuming issue arising from mass updates and major organization changes. The team made a core change in their domain to improve the data model and simplify management/administration in this area of the domain. This resulted in less inconsistencies and a more stable system with minimized risk of data being lost.
And in the end?
Completing the individual sprint goals and the changed state goal after only three sprints were one thing, but equally important was that the whole Ernst Project would make such a big impression that it would be incorporated into the department’s soul, and continue even after the project was completed. In other words, to end with a long- lasting impression that would take form in a changed mindset where it is completely natural for the teams to implement "Ernst-things" in connection with new development initiatives.
Project Ernst will return in…
Overall the project with Ernst was a success since the developers really appreciated the time and confidence given to them. The feeling is that the project gave good dedication to their domain and a consensus between product owners and teams what the cost of having technical debt and postpone handling it.
Who knows, maybe Ernst will be back in the future to ensure he is not forgotten!