Here's a retrospective from the first release of XTrack. Which two or three items would you choose to work on for the next iteration?
Stop
Having uneven distribution of task ownership
Starting the iteration with incomplete requirements
Start
Pairing more, both to write production code and to test
Writing the acceptance tests in advance of the iteration
Having a more regular schedule
Communicating constantly, have regular standups
More knowledge sharing
Continue
Remote pairing
Estimating fairly accurately
Being flexible about changing the requirements when we see things that work better
A1:
This is totally subjective, of course. Our top three would be to stop having uneven distribution of task ownership (this could be reworded as start allocating tasks more evenly), pair more, and have regular standups. We're in two locations, which makes pairing and communication a special challenge.
