Articles from more than two decades of building, rebuilding, debugging, and leading web projects.
Some are practical development notes; others are longer pieces about the decisions and trade‑offs that determine whether modern web platforms hold together in production. You can also browse articles by category.
Choose where React state should live, comparing local state, context, URL state, server data, cache state, mirrored values, and ownership questions.
How to turn Lighthouse findings into a credible engineering roadmap, separating signal from noise and scores from real user impact teams can act on.
Reviewing Claude logs prompts a question: when AI helps us interpret ordinary situations and write personal messages, how much thinking are we handing over?
useTransition and useDeferredValue to Keep React Interfaces ResponsiveuseTransition and useDeferredValue explained with search, filtering, pending states, and how to keep React interfaces responsive under heavy work.
useMemoProfile React re‑renders before adding useMemo, using DevTools, render reasons, prop identity, state placement, context updates, and list strategies.
Stale closures in React explained through renders, handlers, effects, timers, refs, dependency arrays, why callbacks read old state, and how to fix them.
Debugging hydration mismatches in React and Next.js, including common causes, repeatable fixes, and how to stop server and client renders drifting apart.
AI search readiness starts with crawlable pages, rendered HTML, semantic structure, content quality, and safe action boundaries, not protocol‑first thinking.
Measure GEO with honest evidence from Search Console, analytics, query sets, AI‑assisted journeys, attribution patterns, and outcomes instead of fake scores.
Reduce headless CMS image bandwidth and quota risk with a storage mirror: sync changed assets, reuse originals, and keep Next.js image delivery off your CMS.
Lessons from running Next.js Pages Router at scale in 2026, including governance, performance discipline, and why older patterns sometimes still make sense.
How to de‑risk a CMS migration before the real move starts, by testing workflows, content assumptions, integrations, ownership, and technical contracts early.
Retention theatre explained through surveys, recognition schemes, pay, autonomy, trust, workload, management quality, and why strong engineers still leave.
Why transformation programmes fail after the slide deck, covering ownership, governance, operating‑model change, vendor fragmentation, and delivery reality.
Structure technical AEO content for direct answers without FAQ spam, using extractable passages, visible text, schema discipline, and topic depth.
Build an LRU cache in TypeScript with a hash map and doubly linked list, then compare it with a simpler JavaScript Map implementation and O(1) operations.
My view on product engineers and software engineers: how the remit differs, why organisational support matters, and what we need to learn after release.
How to make content easier for generative engines to trust, with clearer claims, stronger authorship, better sourcing, and structure that reduces ambiguity.
Consultancies can unblock stalled programmes, but without ownership transfer and internal capability growth, rescue work becomes recurring dependency.
Deadline debt explained as hidden rework, incidents, technical debt, quality loss, and recovery cost when teams compress quality without recording trade‑offs.
Diagnose Sass colours that survive compilation but break in Next.js production CSS when minification creates invalid mixed‑unit legacy rgb() values.