The Archive

Best practices by year — what we believed, and what happened next

Pre-load all reference data, lookup tables, and configuration into memory during application initialization

Construct SQL queries by concatenating user input directly into query strings

Ship iteratively, skip comprehensive documentation, respond to change over following a plan, value working software over written specs

Ship iteratively, skip comprehensive documentation, respond to change over following a plan, involve stakeholders continuously

Separate SQL structure from user-supplied data using parameterized queries

Defer data loading until the moment it is actually needed, rather than pre-loading at startup

Wrap remote calls in a circuit breaker that fails fast when a downstream service is unhealthy

Break monolithic applications into independently deployable services, each owning a single bounded context

Expose a dedicated HTTP endpoint that reports application health for orchestrators and load balancers