/Engineering
State a render loop reads belongs in a ref
A canvas game where you could never lose. The death screen showed for one frame, then the game quietly restarted. The bug was in the effect deps.
// writing
From the workshop
Page 4 of 8
/EngineeringA real concurrency cap: a worker pool and a circuit breaker
Promise.all is not a concurrency limit, it is a promise to run everything at once. Two lessons on capping work and cutting off a source that keeps failing.
/EngineeringNormalise at the boundary, not everywhere else
My similarity thresholds kept drifting for no reason I could see. The cause was mixing normalised and raw vectors, so cosine was measuring two different things at once.
/EngineeringIsolate the runtime, codegen the contract
Keeping a heavy Python ML service out of the JS install was the right call. Hand-mirroring the request shapes on both sides was not.
/EngineeringWhy I left WordPress (and you might too)
WordPress is a fine tool that quietly became the wrong one for me. A fair, slightly spicy account of why I left, and how to tell if you should too.
/EngineeringThe write that must succeed
A mailing-list signup taught me to separate the write that has to land from the nice-to-have side effect, and that binding names lie on shared accounts.
/EngineeringIdempotent ingestion is a three-part key
The same external id can collide across providers and must never merge across users. Why idempotent third-party ingestion is a composite key, enforced in two places.
/EngineeringGraceful drain on shutdown: the deploy that killed a three-minute job
A routine deploy landed mid-render. The process died, the job stuck 'active', the output half written. The fix was letting SIGTERM finish the work first.
/EngineeringRebuilding my site: from WordPress to Astro on Cloudflare
I tore the old WordPress site down to the studs and rebuilt it as a fast, repo-native Astro project on Cloudflare. Here is the why, the how, and what I would do again.
/EngineeringDeduplicate on meaning, not on an id
My deduplication code ran on every batch and never removed a single duplicate. It was keying on a random id that is unique by construction, so it could only ever match a row against itself.