Skip to main content

Give it a REST

I'm wrapping up a REST layer to the service backend I've been developing for my still-unnamed-employer (find out when we launch, real soon now!). I had never developed a service under the "REST" acronym before, so my boss gave me a crash course, I read some things, I thought I got it. REST, a buzzword in its own right, is like stapling smoke to water when you try to define it. That isn't because its vague, its because most of the people who talk about it don't know what they're talking about.

Maybe I'm one of them and I shouldn't be posting this.

REST is Not:
  • HTTP
  • XML
  • RPC
  • A protocol, format, or even much of a specification
Rest is:
  • An idea(l)
  • Agnostic on just about every specification associated with it
Requests on a REST Service are:
  • Atomic. No request relies on any other made before or after it.
  • Self Authenticating. Every request must include any credentials. See point 1.
  • Self Describing. This is most commonly XML, and sometimes people think it must be, but it can be anything. We use JSON.
Some of the most interesting things I've learned working with a REST service are the things that do not fit the model well. No model fits every need perfectly, and REST doesn't escape that fact, I'm afraid.

In particular, you are not always transfering a state. There is a distinct difference between state transfer and a request to perform some operation upon a state. Unfortunately, any ways around some of the problems posed are directly rejected by the rules of REST.

For example, say you want to provide as a service a simple counter. You expose PUT on /counter/foobar to register a new counter, and then GET on /counter/foobar will provide the current level of the counter. Following the rules of REST, how do you provide an interface to safely increment such a counter? We can not perform a GET and a PUT, because it violates that each request be self contained, and it will break when any other client of the service is incrementing at the same time. We need a single operation to alter the state, without performing a state transfer.

The best thing you can do is use POST on a resource, and transfer a request to increment. It seems to violate the tenents of REST that the resource you POST will not actually reside at some permenant location, as they are throw-away requests. You either have to live with a not-exactly-REST interface (but, isn't that it works the important thing?) or actually keep requests around for some time. Maybe put them at some location, where they can be checked for review of their status.

I don't know if this is helpful to anyone else writing REST services, but the information around isn't always accurate, so why should I worry if I am?

Comments

Andrew Dalke said…
""""It seems to violate the tenents of REST that the resource you POST will not actually reside at some permenant location, as they are throw-away requests. """

That's not a tenent of REST. Think of submitting a blog entry. You POST to a resource which creates another URI (for the entry itself) and updates the main page.

For a counter you would not PUT at all, except perhaps to (re)set a counter. You could POST to a counter and have it increment by one, or GET from the counter to see its current state. Or use two URLs, one for each.

POST is a catch-all verb which has no explicit limitations on what it can do. GET should be side-effect free, PUT should only modify the resource PUT'ed to, and DELETE should only delete the resource PUT'ed do.

They can have side effect, eg, deleting an object likely means a resource listing all items in a collection gets updated. But the side effect should fit with the action.

POST, though, is free to do anything. Hence proxies and caches can't make any assumptions about its effect.
Anonymous said…
Agreed... PUT is when you are placing (uploading) a resource to a pre-known location. Thus, /foobar/counter/1 would reset the counter to one. POSTing to a counter would increment it by one. If you were looking for a truly atomic "increment only if the current counter is less than 5" then you would use POST again.

Here's the basic difference: with RPC (i.e., XML-RPC, SOAP, etc), you call /getperson?name=jamie while with REST you'd call /person/jamie with a GET command.

In other words, with REST you call or create a resource -- a database record, an object in the OOP sense, the model -- with one of three basic VERBS, GET, PUT, or DELETE. (POST can be for RPC-like behavior when you don't actually know what the resource might be called.)

I.e., instead of calling a FUNCTION or METHOD and pass what object you want to call as a parameter, you instead call the remote object and pass the function (GET,PUT, DELETE) you want as the parameter.

It's actually not too hard, just requires an adjustment in thinking, but I agree -- most people that think they know what REST is, don't!

Popular posts from this blog

CARDIAC: The Cardboard Computer

I am just so excited about this. CARDIAC. The Cardboard Computer. How cool is that? This piece of history is amazing and better than that: it is extremely accessible. This fantastic design was built in 1969 by David Hagelbarger at Bell Labs to explain what computers were to those who would otherwise have no exposure to them. Miraculously, the CARDIAC (CARDboard Interactive Aid to Computation) was able to actually function as a slow and rudimentary computer.  One of the most fascinating aspects of this gem is that at the time of its publication the scope it was able to demonstrate was actually useful in explaining what a computer was. Could you imagine trying to explain computers today with anything close to the CARDIAC? It had 100 memory locations and only ten instructions. The memory held signed 3-digit numbers (-999 through 999) and instructions could be encoded such that the first digit was the instruction and the second two digits were the address of memory to operate on

Statement Functions

At a small suggestion in #python, I wrote up a simple module that allows the use of many python statements in places requiring statements. This post serves as the announcement and documentation. You can find the release here . The pattern is the statement's keyword appended with a single underscore, so the first, of course, is print_. The example writes 'some+text' to an IOString for a URL query string. This mostly follows what it seems the print function will be in py3k. print_("some", "text", outfile=query_iostring, sep="+", end="") An obvious second choice was to wrap if statements. They take a condition value, and expect a truth value or callback an an optional else value or callback. Values and callbacks are named if_true, cb_true, if_false, and cb_false. if_(raw_input("Continue?")=="Y", cb_true=play_game, cb_false=quit) Of course, often your else might be an error case, so raising an exception could be useful

How To Teach Software Development

How To Teach Software Development Introduction Developers Quality Control Motivation Execution Businesses Students Schools Education is broken. Education about software development is even more broken. It is a sad observation of the industry from my eyes. I come to see good developers from what should be great educations as survivors, more than anything. Do they get a headstart from their education or do they overcome it? This is the first part in a series on software education. I want to open a discussion here. Please comment if you have thoughts. Blog about it, yourself. Write about how you disagree with me. Write more if you don't. We have a troubled industry. We care enough to do something about it. We hark on the bad developers the way people used to point at freak shows, but we only hurt ourselves but not improving the situation. We have to deal with their bad code. We are the twenty percent and we can't talk to the eighty percent, by definition, so we need to impro