Steve, I think you're quibbling, here. Yes -- Fielding "documented" the Web's architecture after the fact. But nonetheless, REST describes the Web's architecture.
Anne On 12/19/06, Steve Jones <[EMAIL PROTECTED]> wrote:
On 19/12/06, Anne Thomas Manes <[EMAIL PROTECTED] <atmanes%40gmail.com>> wrote: > > > > > > > On 12/18/06, Steve Jones <[EMAIL PROTECTED]<jones.steveg%40gmail.com>> wrote: > > > > On 18/12/06, Anne Thomas Manes <[EMAIL PROTECTED]<atmanes%40gmail.com>> wrote: > > > > > > NOTE: REST *is* a disruptive innovation. I can't think of too many other innovations that have been as disruptive as the hypermedia architecture that has enabled the Web. I'm less convinced that the application of REST to application integration will be quite as disruptive. > > > > Completely agree, but REST being "the web" is a big leap as well. > > I've done a bunch of Web projects, and know others who have done even > > more. None of us followed the principles of the REST paper. I'm not > > disagreeing with the 2nd bit Anne, but I think the first bit is one of > > those bold claims that don't stack up. So was WWW and HTTP > > disruptive... hell yes. But did this REST architectural approach > > enable that, or did we all just get on with using a browser as a thin > > client... I'd say the later. > > REST defines the core architecture of the Web -- URIs, resources, > representations, self-describing media types, the idempotent GET > operation that enables caching, etc. Although many applications don't > follow the REST principles to the letter, the Web would not work if it > weren't for REST. Which came first the Web or the REST paper? The answer is of course the Web, REST does a good job of documenting much of the mechanics of how the Web works, but if the REST paper didn't exist the Web would still have been as successful. The REST paper does a good job of documenting what existed and then suggesting new ways in which it could be used. It documents pre-existing elements such as the return codes and the operations and suggests how an increase in formalism (e.g. GET = idempotent) could help applications perform. It also then further proposes how server->server collaborations could be built by using the same hyperlink approach that was previously used for browsers. So REST doesn't _define_ the way the Web works, it _documents_ and proposes how it could _also_ be used. There is a big difference between the two. TBL defined the core of the Web, lets not go all historical revisionism just yet.
