GraphQL let clients write their own queries because no server team could predict Facebook's graph, and a team that locks it down to persisted queries has already shown it doesn't have that problem.
Read →Why deciding which customers are owed a feature doesn't belong in a flag service, when a flag can stand in anyway, and how client apps should get the answer.
Read →Why business rules drift into API gateways, what each one loses when it lands there, and a one-question test for what belongs.
Read →Ordered routes through the site toward a goal, from foundations and arguments up to advanced practice, with a line on why each step comes where it does.
Browse paths →Search the full text of every guide, post, resource, and case study: use the magnifier in the header, or press /.
Two guides on coding agents: Beads keeps their work from getting lost between sessions, and Gas City runs many of them reliably.
Why wide update bodies can't tell an omitted field from a cleared one, and how small, replaceable write resources avoid the problem.
Read →A Claude Code skill that turns a ticket into a self-contained plan, tags every assumption with its source, and proves them before any code is written.
Read →A portable Claude skill that lints prose for AI tells and filler, then runs a self-review for what a grep can't catch.
Read →Last updated:
