The expectation that senior developers should be able to work independently is killing the team's ability to collaborate
The more they ask for help the less they feel they are meeting that expectation.
»Measuring Accelerate metrics in XP teams is as redundant
as measuring code coverage for a team doing TDD.
»Reducing transaction cost is necessary, but not a sufficient condition for reducing the batch size
Very often, for a mindset that emerged inside environments with big batches, it takes increasing the holding cost of inventory to start changing the belief system that was created as a workaround for the limitations of the system and served as its governing constraint.
»You might think that small PRs are good
because you see more feedback and quality being built in than on big PRs (‘LGTM + Approve syndrome’).
»Assume you're the one being wrong
»Bosses say “Does that make sense” … to get compliance.
Leaders ask “What didn’t make sense?” to invite discussion.
Planning too soon?
The higher the utilization in the system, the lower the availability and responsiveness, and the earlier you feel an urge to start with work in order to plan and coordinate it.
»Because there is no wait time
the cost of code review while co-creating (pairing/mobbing) is zero.
»When evaluating the slowness of the review process
you have to take into account how long did it take you to make a change (processing time) not just how long it took to get a review (wait time).
»If it takes me 1 minute to rename a method
and 3 hours to get a review for the change, I’m either not going to rename the method or I’m going to rename it as part of some later change.
»One thing about being specific
in the design of methods and messages is that it hugely reduces coupling and enables faster current and future changes.
»