Tech and case studies on organizational management, team management, and engineering efficiency
The article argues that in early development stages, a small amount of code duplication should be tolerated, warning against premature DRY abstractions that can lead to over‑engineered designs, and advises waiting for genuine common patterns before introducing shared abstractions.
Google's research demonstrates that a machine‑learning model trained on build logs and code snapshots can automatically suggest safe, high‑quality fixes for broken builds, boosting developer productivity by about two percent without introducing detectable security risks.
The article reflects on a technical documentation guide, discusses Tencent's training series, highlights the challenges developers face in finding time for documentation and reviews, and argues that providing sufficient resources is essential for effective documentation practices.
The article reviews the history of China's IT industry over the past eighteen years, exposing how government policies, CMM certifications, and superficial agile practices have led to chronic software quality and delivery problems, and argues that only solid fundamentals such as extreme programming, TDD, and deliberate practice can truly improve software development.
The article explains why Google created an official Code Health Group, how it promotes readable, maintainable code through internal documentation, consulting, and tooling, and why such organized efforts often fail in other companies lacking formal support and sustained leadership.
This article recounts how Google engineer Laurent Le Brun and the Bazel team designed, built, and rolled out Buildifier—a code‑formatting tool that automatically reformatted over one hundred thousand BUILD files—detailing the technical challenges, testing strategies, large‑scale deployment, and the resulting improvements in code consistency and developer productivity.
This article examines Google’s two‑decade testing culture, detailing the iconic “toilet test” posters, the development of a testing certification program, the shift toward developer‑driven automated testing pyramids, the challenges of scaling massive test suites, and the evolving role of test engineers within the broader quality engineering framework.
The article recounts how the Bugzilla project, originally tied to a single proprietary database (OldDB), was refactored to support multiple databases—including NewDB—through a comprehensive redesign that involved breaking the work into smaller, well‑tested tasks, improving maintainability and expanding support to four database systems.
This article examines the practical challenges of using large language models in software development, including handling long contexts, cross‑file editing, bug‑fixing evaluation methods, and presents benchmark results from SWE‑Bench and its Lite subset to assess model capabilities.
The article explains Codeium’s two new metrics—Characters per Opportunity (CPO) and Percentage of Code Written (PCW)—how they are calculated, why they better capture the true value of AI‑powered code completion compared to traditional benchmarks, and how they guide product decisions.
