REST is a resource-oriented architectural style for networked interfaces. It represents each entity at an endpoint and uses familiar web semantics to access those resources. For example, GET /users/123 returns a user, while GET /users/123/posts returns related posts. The model is simple, widely understood, and works naturally with HTTP caching and load balancing. section 5.3.3
The main performance risk appears when a client needs related data. A REST client may make one request for each resource, wait for sequential responses, and receive fields it does not use. Each additional round trip adds latency, with slower networks and mobile clients affected most. REST therefore trades a straightforward endpoint model for potentially more requests and larger cumulative response costs. section 5.3.3
Evaluate a REST design by checking:
- How many resource requests a common screen or workflow requires. section 5.3.3
- Whether responses contain fields that the client does not need. section 5.3.3
- Whether HTTP caching and load balancing can support the resource access pattern. section 5.3.3
Pull-request details on mobile
A pull-request display required separate requests for the pull request, commits, comments, and reviews. Sequential calls delayed complete information for mobile users, while each response also carried fields the client did not need. This illustrates REST's resource clarity and the round-trip costs of assembling related data. section 5.3.3





























