02 tools
Web
Slugs and HTTP status codes.
The unglamorous parts of shipping a page
Two things every site eventually needs and nobody enjoys doing by hand: turning a headline into a URL that survives being pasted into a chat window, and working out which of the forty HTTP status codes actually applies to the situation in front of you. Both are small, both are easy to get subtly wrong, and both are hard to fix later — one because links have already been shared, the other because clients have already been written against your API.
Slugs are permanent in practice
Once a URL is indexed, changing it costs you whatever pointed at it. Serve a 301 and keep it, or accept the loss. Lowercase, hyphens for word boundaries, no dates, and only as many words as the meaning needs.
Status codes are a contract
The code is the first thing every client, cache and crawler reads. A 200 carrying an error hides failures; a 401 where a 403 belongs sends well-behaved clients into an infinite re-authentication loop. Getting these right is cheaper than debugging them later.