tl;dr:“How do you balance risk, novelty, production, growth, short-term certainty, and long-term viability? I learned a simple rule that has been useful to me and is often cited by my students as a key lesson from coaching: (1) 80% of your time goes to low-risk / reasonable-reward work, (2) 15% of your time goes to related high-risk / high-reward work and (3) 5% of your time goes to satisfying your own curiosity with no thought of reward.
tl;dr:Kent explains what Snapshot Testing is and how it scores on the test desiderata - a list of 12 desirable properties of tests. This list is a useful framework for evaluating different types of tests.
tl;dr:“Easy-to-test software is "controllable". Testers can cheaply and accurately simulate the contexts in which the software needs to run. Two contradictory patterns help achieve controllability: making parameters more concrete and more abstract. This apparent contradiction resolves when looked at from a broader perspective.”
tl;dr:“In fact, I believe that our skills as software developers are more valuable than ever before. While AI tools like ChatGPT can certainly automate routine tasks and help us be more efficient, they can never replace the human creativity and expertise that is essential to delivering high-quality software products.” Kent expands on this.
tl;dr:Blame is the opposite of responsibility — it’s someone with power pushing consequences away from themselves & onto someone with less power. The word “accountable” is often used as a proxy for “blame” - it’s not a relationship building strategy but used as a “relationship decaying” strategy. Kent discusses how this shows up amongst management.
tl;dr:At Kent's company Gusto, the engineering team switched from private to public engineering levels and, as part of the transition, Kent wanted to emphasize to senior engineers and managers that power should not be misused. Here is Kent's ways in which to exercise "power advantages" & experience power disadvantages.
tl;dr:As an idea, product or company grows, "value-maximizing behavior" changes dramatically and comes in three phases: (1) Explore - where companies try small experiments. (2) Expand, as an experiments takes off, bottlenecks are identified and tackled. (3) Extract, these projects can increase revenue or decrease costs.
tl;dr:Latency is “the time interval between a stimulus and its response." Throughput is “the rate at which a system achieves its goal”. Kent discusses the relationship of these two in architectural decisions, and when to optimize which.
Click the link in this tweet if you're paywalled.
tl;dr:Two experiments testing different workflows were conducted. One involved making small changes, instantly deployed. The other, automatically committed code that passes its tests, deleting what fails. Both should have gone wrong but didn’t.