code quality

Bilingual

Beyond "Clean Code": Why Your Comments Matter
2 months ago
- Comments are necessary to express intent and 'why' information that logic alone cannot convey.
- Uncle Bob's claim that 'comments are always failures' is too extreme and can be harmful.
- Logic only shows what the code does, not what the programmer intended or the reasoning behind decisions.
- Comments should be used to communicate naturally without contorting the logic.
- Objections about comments lying or going stale are less problematic than having no comments at all.
- Comments improve both good and bad code; they are not a crutch for poor code.
- The author challenges critics to eliminate comments without losing information or contorting logic.
Cull your dependencies
2 months ago
- - The Log4J vulnerability serves as a cautionary tale, demonstrating how a widely used logging library (168,000 lines of code) can introduce critical security risks.
- - Dependencies accumulate enormous amounts of code (often >1M lines) with an estimated 15–50 bugs per 1,000 lines, leading to thousands of potential issues in imported code.
- - Developers often import full libraries to avoid writing small utility functions, adding unnecessary bloat and unknown code quality.
- - There is a false assumption that packages from official package managers are inherently high quality and secure, despite being maintained by volunteers with no required checks.
- - Dependencies become entrenched and proliferate due to assumed best practices, making removal difficult and causing vulnerabilities to silently accumulate.
- - Poor tooling for detecting unused dependencies leads to stale packages persisting as bloat.
- - Proposed rules: avoid adding dependencies, require explicit justification and logging for each addition, and standardize on a single utility package.
<antirez>
2 months ago
- The best code is often written when the programmer is supposed to be doing something else, as seen in open source projects.
- Open source software reflects a programmer's passion, joy, or anger, driving them to pursue perfection.
- User demands in open source are not about money; they reflect shared care for software quality.
- Contributors help improve the software; ignoring them is an option, but their input stems from shared concern.
- The creator has the right to control design and refuse contributions that don't fit the vision.
- Treating user interactions as merely about money is a mistake; both parties value quality.
- Editors note that upset OSS developers might find their work resembles office work.
- The paradox highlights that OSS writers care more about free code than paid work.
Anyone else get a vague GitHub shakedown notice?
3 months ago
- Greg Troxel received an email from GitHub about billing for Code Quality, moving to general availability in July 2026.
- He suspects the email is genuine but is confused as he has not signed up for paid plans and is not a GitHub Sponsor.
- Greg questions if QGIS's use of Code Quality might have triggered the email and asks if others received similar notices or are concerned about unexpected bills.
Why you should not have Managers in your class name
3 months ago
- Naming classes and methods correctly is crucial as it influences code structure, future thinking, refactoring, and extension, with poor naming leading to technical debt.
- Using '-er' suffixes like 'Manager' often indicates poor design; for example, 'ConnectionManager' became complex with multiple responsibilities, whereas 'Connection' would have been clearer and simpler.
- Choosing precise names helps maintain clean code, simplifies extension and refactoring, and reduces complexity and bugs, so spending extra time on naming is recommended.
Zero-defects code: the prescient Microsoft memo from 1989
3 months ago
- Microsoft's 1989 Zero Defects memo advocated for principles like continuous integration and test-driven development long before they became mainstream.
- The memo highlighted problems in Microsoft's culture, such as rewarding developers for checking in incomplete code and treating schedules as more important than quality.
- Chris Mason argued that writing perfect, zero-defects code was achievable and emphasized the need for a mindset shift to prioritize quality over speed.
- Key solutions proposed included having a working, shippable product daily, fixing bugs immediately, and shifting testing to quality assurance rather than bug-finding.
- The memo suggested practical techniques like code reviews, writing tests before debugging, and using asserts to verify assumptions, aiming to reduce bugs and improve predictability.
- Leadership committed to supporting realistic schedules and pushing back against pressures to cut corners, emphasizing that taking minor heat early was better than major issues later.
The Short Leash AI Coding Method for Beating Fable
3 months ago
- The post discusses using AI agents to write high-quality software, emphasizing a method suitable for expert developers who want to enhance performance without compromising quality.
- Current AI agent approaches, like those promoted by YouTubers, often lead to low-quality code and lack of understanding, as AI can go off track, especially in niche areas with limited training data.
- The "Short Leash" method is introduced, which involves active developer involvement, including planning, reviewing diffs, denying permissions when needed, and committing after subtasks to maintain control and quality.
- For AI reviews, combining human and AI review is recommended, with AI acting like a linter and humans catching higher-level issues. PRs should disclose AI usage and be thoroughly reviewed by the author before submission.

code quality

Beyond "Clean Code": Why Your Comments Matter

Cull your dependencies

<antirez>

Anyone else get a vague GitHub shakedown notice?

Why you should not have Managers in your class name

Zero-defects code: the prescient Microsoft memo from 1989

The Short Leash AI Coding Method for Beating Fable