UUIDs are practical identifiers, not guarantees or secrets

A UUID lowers coordination requirements; it does not remove the need for a database uniqueness constraint. Random identifiers can also fragment a heavily written ordered index, so systems with that workload may prefer a time-ordered identifier while keeping the public ID opaque. Validate the canonical representation at boundaries and store it in a native UUID type when available rather than an oversized arbitrary string. Never infer creation time, authorization or trust from an identifier unless its format and security model explicitly guarantee that property.