We need fewer assumptions to test
with each deployment (smaller changes), not more delayed feedback (more gates, more branches, etc.).
»with each deployment (smaller changes), not more delayed feedback (more gates, more branches, etc.).
»(same as with async code reviews) is that the author knows when and what kind of feedback they need.
»Sure, just be aware that your sense for determining “when it makes sense” was shaped by the opposite way of working (work in isolation) which likely optimizes for having hard time finding scenarios when pair/mob makes sense.
»(say that a simple, yet good enough proxy for that is halving the size of a PR, or halving the time taken to create a PR) but you don’t halve the wait time to review it, you’re losing throughput of number of changes you’re able to push through the system of work in a unit of time.
»is that the system doesn’t give a damn that you’re mad at it because you’re getting outcomes opposite to the ones you’d like to get.
»I often hear: “5 min rotation doesn’t seem enough, I think 10 min would be more appropriate”.
»The more we speculate upfront about what the design should be, the less likely that we’ll build the contextual awareness muscle needed to look at the existing design and ask “what is it trying to tell me that I’m not hearing?”.
»The idea of “on time” as a project health metric is deeply flawed. People get satisfied and excited about it when a project is “on time”, but it essentially communicates overconfidence of people and organizations trying to fit the inherently complex reality into a number they came up with. The truth is: the reality doesn’t give a damn about what we think of it.
»One of the big benefits of slicing, that I don’t get to hear at all, is that the way we sequence slices by value, as a byproduct also serves as a guide on where are the fracture planes or seams for gracefully degrading business capabilities when things go south in production.
»From the “right leverage point, wrong direction” series (reference to Jay Forrester), the majority of teams I worked with when faced with high batch transaction cost relative to batch size, instead of reducing the transaction cost they increase the batch size.
»