REST in Practice

Foreword

FROM THE VERY START OF WHEN I GOT INVOLVED IN COMPUTING, there’s been the desire to have software systems designed as components that can be freely combined. The wide-scale connectivity of the Internet fueled this desire, and added the desire to have components operate over networks that introduce issues of latency and unknown reliability. In this world many systems have been tried, and many have failed—usually with a whimper.

A great example of success is the World Wide Web. Its success has penetrated both business operations and popular culture. It provides opportunities for people to pull together information from many sources, with hardly any prearranged collaboration—and at a global scale.

The Web, as we currently know it, isn’t the be-all and end-all of computing, but many people believe it offers an important lesson on how to construct systems of networked components. Many people take advantage of its protocol, HTTP, to connect systems. But some people think we should go further, using HTTP not as a convenient tunnel, but to embrace the way the Web works as a foundation for systems collaboration.

This thinking gathers together under the name of “REST.” It refers to Roy Fielding’s PhD thesis, which is far more often referred to than it is read. There is a growing notion that following the principles of REST offers a fruitful path to making networked components work, one that is built upon the success of the Web itself.

That vision is attractive, but there is much to be ...