CHAPTER 5: Representational State Transfer (REST) | Hacker News Reader

CHAPTER 5: Representational State Transfer (REST)

ics.uci.edu

34 pointscollint12 comments

I've about had it reading discussion after discussion on the topic of REST between people who haven't read the damned paper.

Here is my takeaway:

There isn't much to do with URLs in here. One URL, One resource (collection counts as a resource). Talk about URLs all you want, there can bee good resource identifiers and bad ones. But REST doesn't care as long as you don't break the golden rule.

REST is an RPC system. You call remote procedures, GET, PUT, DELETE etc. on remote resources. You are probably not smart enough to do better than what naturally emerged as HTTP. Behemoths and industries have tried. But HTTP won, and it will win again.

Sorry that's wrong; neither REST as a style or HTTP as an implementation are RPC. One of the key components of REST is Layering which is made possible because REST is distinct from RPC. This is covered in the dissertation:

http://www.ics.uci.edu/~fielding/pubs/dissertation/evaluatio...

6.5.2 HTTP is not RPC

People often mistakenly refer to HTTP as a remote procedure call (RPC) [23] mechanism simply because it involves requests and responses. What distinguishes RPC from other forms of network-based application communication is the notion of invoking a procedure on the remote machine, wherein the protocol identifies the procedure and passes it a fixed set of parameters, and then waits for the answer to be supplied within a return message using the same interface. Remote method invocation (RMI) is similar, except that the procedure is identified as an {object, method} tuple rather than a service procedure. Brokered RMI adds name service indirection and a few other tricks, but the interface is basically the same.

What distinguishes HTTP from RPC isn't the syntax. It isn't even the different characteristics gained from using a stream as a parameter, though that helps to explain why existing RPC mechanisms were not usable for the Web. What makes HTTP significantly different from RPC is that the requests are directed to resources using a generic interface with standard semantics that can be interpreted by intermediaries almost as well as by the machines that originate services. The result is an application that allows for layers of transformation and indirection that are independent of the information origin, which is very useful for an Internet-scale, multi-organization, anarchically scalable information system. RPC mechanisms, in contrast, are defined in terms of language APIs, not network-based applications.

Shorter RTF: The client might GET from a cache, so obviously GET isn't a procedure running on the remote server.

Yeah, caching is a type of layering

Well color me stupid. I suppose I'm going to have to be more careful about what I call RPC.

You managed to read it all? I printed his dissertation out once and fell asleep at the beginning. If this happened to anyone else you can read a somewhat abridged version in his ACM paper: http://www.ics.uci.edu/~taylor/documents/2002-REST-TOIT.pdf

On another note, I remember reading his "musings" and getting the impression that he was mean and arrogant. I met him last year and he seemed like a nice guy. I was still careful to call my serial device HTTP resource server "RESTish" out of fear that I probably violated some principle somewhere.

So in reading that chapter it seems like REST can be boiled down to these statements:

1) Client-server

2) Stateless

3) Cacheable or noncacheable constraints

4) Uniform interface

5) Abstracted so that you can only see the layers directly in front of you.

6) Client side code, e.g., Javascript

7) (Potentially) Late binding of reference to representation

8) "The connector interface is similar to procedural invocation, but with important differences in the passing of parameters and results. The in-parameters consist of request control data, a resource identifier indicating the target of the request, and an optional representation. The out-parameters consist of response control data, optional resource metadata, and an optional representation."

But a lot of what is described as REST imposes additional constraints that I don't see in this description. Did I miss something?

I tend to agree with you, but don't just take it from me: http://tech.groups.yahoo.com/group/rest-discuss/message/6735

I don't think it's fair to say the original dissertation is the final word on REST, but I do think almost all of the discussion on REST is way off target. It feels like almost all the discussion is about the construction of URLs.

REST has become a buzzword, people use it just to project credibility.

I have a lot of respect for him, but it's pretty disappointing to see him so opposed to doing things that would bring more clarity to those trying to understand his work. This is at least the third quote I've seen of him blowing off the idea of giving examples or more documentation. It's further disappointing to see "that document is my dissertation" juxtaposed with "I sometimes need to expand on the terse explanations provided in the dissertation, because I was in a hurry when I wrote it".

One (additional?) constraint that perhaps doesn't fall nicely into one of those categories is "Hypermedia as the Engine of Application State" (HATEOAS), which was emphasized (though not with those exact words) by Roy Fielding on his blog:

http://roy.gbiv.com/untangled/2008/rest-apis-must-be-hyperte...

It's part of the uniform interface constraint

http://www.ics.uci.edu/~fielding/pubs/dissertation/rest_arch...

Shouldn't that be RST?

Reply on news.ycombinator.com