Architecture of the Web

• biggest info system ever
• success because flexible & incremental
  • layers on top of already functional layers

REST (Representational State Transfer)

• key principles

  • all sources are resources, uniquely accessible via URI

  • clients & servers need to know:

    1. URI of resource
    2. Allowed actions (GET, POST, HEAD…)
    3. Allowed representations (text/html or others)

  • client doesn’t need to kno how server generates representation

  • server doesn’t need to know how client renders content

  • neither have to be aware of intermediate proxies

  • there is no HTTP communication state (response doesn’t depend on previous request, no side effects)

    • delete/put are unsafe but idempotent
    • get and similar are safe
    • post needs warning, can happen multiple times
• typical HTTP requests
  • GET data
  • POST (e.g. submit form)
  • PUT (update info)
  • DELETE