For a while now, I’ve held the belief that the biggest reason people get the whole “REST” thing wrong is because they are looking at Roy Fielding’s doctoral dissertation – the paper that coins the term “REST” - as a prescription for how to design APIs.
15 Oct 2016
10 Oct 2016
In response to my recent post on Swagger and REST, I was asked the following on Twitter:
07 Oct 2016
If you’ve spent much time with me, you’ve undoubtedly heard me ramble on at length about linked data. And in those conversations, you’ve likely heard me say something to the effect of “linked data is REST”. However, I haven’t really spent much time talking about REST by itself - especially considering the amount of importance heaped on it by proponents of the “API Economy”. I’ve focused my attentions elsewhere primarily because as an architectural style, REST isn’t something that a team can just go and implement. Rather, REST describes (in the form of constraints) the properties of the World Wide Web.
16 Sep 2016
While I love the simplicity of Jekyll for generating static Web sites from markup documents, the fact that Jekyll is built on Ruby and its respective ecosystem has been giving me increasing frustration. The reason for the growing frustration comes down to this:
15 Sep 2016
I’ve spent some time talking with my teams recently about REST from the perspective of an API, and also how APIs built through the lens of technologies like Swagger go against some of the most important principles of REST. As I’ve reflected more over conversations, there is one even more fundamental thing that tends to get lost in the conversation. And that is that REST doesn’t describe APIs. REST describes the architectural characteristics of an entire system, which includes all of the different components of that system. Trying to make a systems-level claim based on one component, or even one layer, of that system simply doesn’t work.