Work studies
Short, one-week deep dives. Each one either takes apart how a well-known app works, or redesigns an older tool people still rely on. Each ends with a small working piece for a planned feature in one of my projects.
- Nov 2026How VS Code’s terminal keeps up with a flood of outputVS Code’s terminal, which is built on xterm.js, stays smooth even when a program floods it with output. I’ll dig into how, then spend a week building a terminal pane that stays smooth and keeps its memory use under a hard ceiling.
- Nov 2026How Google keeps automated code review worth readingGoogle runs many automated checks on every code review, and anyone can mark a finding as not useful. Checks that annoy people get fixed or switched off. I’ll study how that system, Tricorder, keeps the noise down, then spend a week building a triage step for AI reviewers.
- Nov 2026How Clippy and ESLint decide which fixes are safe to applyRust’s Clippy and JavaScript’s ESLint fix some problems on their own and only suggest fixes for others. I want to understand which fixes a tool may apply on its own and why, then spend a week building that split into a checker for interface code.
- Nov 2026How Deno and macOS sandboxes make a program read-only by forceDeno won’t let a script touch files or the network unless it’s been given permission, and macOS can wall a process off with a sandbox profile. I’ll look at how both enforce it, then spend a week building a check that proves a command-line tool can’t write.
- Nov 2026How axe-core keeps false alarms rare enough to trustaxe-core is the accessibility checker behind many browsers’ developer tools. It’s built to avoid false alarms, and when it can’t tell, it says so instead of guessing. I’ll study how its rules are designed and tested, then spend a week building a way to measure how often a design rule is right.
- Dec 2026How chess and baseball analytics tell a good decision from a lucky resultLichess grades every move by how good it was, whatever happened next. Baseball Savant’s expected stats show how a hit usually turns out, apart from where this one landed. I’ll study how both separate the decision from the result, then spend a week designing a way to explain a single chance outcome.
- Dec 2026How merge queues land parallel changes without breaking the main branchTwo changes can each pass their tests and still break the main branch together. GitHub’s merge queue, bors and OpenStack’s Zuul solve this by testing changes in the order they’ll land. I’ll study how, then spend a week building a small queue for work done in parallel.