An atomic data type identifies the kind of single value a data element contains, such as a string, number, or boolean. Choosing the type affects how consumers interpret and process API data. section 7.4 section 8.5
Choose atomic types as part of a broader data-modeling workflow:
- List the concept’s elements before debating final names or types. section 5.2.5
- Group related elements into objects or arrays, then choose each element’s type and whether it is required. section 5.2.5
- Prefer correctly named booleans, and review strings and numbers for human readability and interoperability. section 8.5
A type choice changes consumer effort
A washing machine value such as {"ec": 50400} may technically communicate an end-of-cycle time, but consumers must decode and calculate it. A human-readable value such as {"endOfCycle": "2:35 pm"} communicates the meaning directly. The same concern applies when modeling a Product resource: productReference can be an integer, while supplier can be an object. section 8.0 section 5.1.3
Figure 8.11 Contrasting non-user-friendly and user-friendly data types and formats
Compare user-friendly and non-user-friendly atomic data types and formats. section 8.5
