[tvjames]

Useful article for someone needing help in revising a URI naming scheme, but beyond that could end up being harmful to that person's greater understanding of building "RESTful" systems.

For example, in the Idempotency section, the author states that for GET requests "no change in application state should occur" which is good, but then also states "the response should always be the same" which is incorrect. There's nothing that says the response cant change, if the resource has changed, but the constraint is that it cant cant _as a result_ of the GET request. This is an important distinction.

DELETEing a resource twice, the second delete should be a NOOP, not a 404.

There's some excellent NDC Oslo & London talks covering more in-depth RESTful topics that I'd recommend checking out if the content of the article is an eye opener for you.

https://vimeo.com/131631886 https://vimeo.com/131196782