Stacked Diffs and Code Review

The State of Software Engineering 🔨 — with Kent Beck

Topic 6 of 7

Published Feb 15, 2024 · Clip 33:46–42:44 (8:58)

Responding to a chat question about stacked diffs, Kent offers his 'spicy take': stacked diffs are a reasonable response to an unreasonable system. He likens them to speculative execution in CPUs, questions the value of blocking asynchronous code reviews, and proposes ideas like non-blocking asynchronous review via a code-change 'newsfeed' ticker. The host mentions the Ship/Show/Ask framework, and Kent connects it to the distinction between reversible structure changes and mostly irreversible behavior changes from his book Tidy First. He argues programmers should separate structure and behavior changes in their diffs.