I was wondering if there is some kind of standard way to describe a
RESTful API? Has somebody done this already and can give examples?
I myself have some ideas:
* You can use UML class diagrams to describe the representations
resources and/or schemas (but these are harder to read and can be
generated from the diagram)
* You can use UML sequence diagrams to describe dynamic behavior. So
how the API is supposed to be used.
* How do we describe the possible operations on resources?
It is not always wanted to change resources or maybe custom errors
must be returned.
PS We use the RESTful protocol for inter-system communication, we do
NOT use it for WorldWideWeb access.
Hello, I was wondering if there is some kind of standard way to describe a RESTful API? Has somebody done this already and can give examples? I myself have...
... Whether the protocol is for internal or external consumption is moot; a protocol is a protocol is a protocol. There is no standard way, but there are ...
Hi all, I profit that Chuck talk about WADL to ask if there is a place for a description language in the REST architectural style ? One REST principle is the...
Hi Benoit, ... I think the key thing is that "hypertext as the engine of state transition information". The goal is to use HTML (or XML) as the mechanism for...
... This is a contradiction in terms. REST is defined by two properties: 1. Server state is exposed as a set of resources that have a uniform interface and are...
... Aristotle, I think that you are right on the money. I am putting together a new talk for the user community here in the UK and that has led me to think a...
... Sigh. This is the problem I discussed when I first started following the rest-discuss list; people's insistence on interpretting the URI opacity axiom to...
... No one said you *should* create obtuse URI spaces; the point is that the layout of your URI space has no effect on whether your application is RESTful. The...
... Agreed. But from past experience many people *interpret* that it means you should create obtuse URLs. It's called dogma, and is sadly appears around any...
... So just exactly what do you propose? -- -Mike Schinkel http://www.mikeschinkel.com/blogs/ http://www.welldesignedurls.org http://atlanta-web.org -...
You could write in your blog about when it is appropriate to use non-obtuse URLs and when it's appropriate to generate URLs on the client and what situations a...
... To my mind it's always appropriate to use non-obtuse URIs. However, degrees of obtuseness are often in the eyes of the beholder. This is one of the reasons...
... Two replies to that: 1.) Can I assume you haven't been reading my URL blog? (of course, I assume someone hasn't until they tell me otherwise.) There I have...
... OK, I'll bite: The argument you make here seems to be for hackable, human-readable URLs. I prefer them myself to meaningless, opaque strings. But I don't...
Not quite. OO systems target humans and therefore benefit from semantics encoded into names. REST targets machines and relies on discovery through contextual...
... 1. Humans tend to come into the process at some point, even when the intention is that they don't. 2. Relative URI-reference syntax depends on a structure...
... Earlier in the thread I gave a hypothetical example for educational purposes, see http://tech.groups.yahoo.com/group/rest-discuss/message/8615 This was in...
... semantics (hypermedia). I thought the web at large *was* REST, so more than just for machines. At least that's what Roy Fielding said about eight months...
... Well... it's not all REST is it? I mean there are plenty of non-RESTfull sites out there which clearly can't be RESTfull. The thing is... those tend to be...
... Web sites which 'clearly can't be restful?' I think this begs the question as to whether you can provide some examples... You mention that they 'tend to be...
... No. I don't mean that. Though that would be another example. I mean applications that don't conform at all to REST. There are quite a few. Hotmail for...
... or gMail, or Yahoo! mail ... Virtually every e-commerce site ... It's probably easier to list the RESTful ones! ... Not necessarily. It is relatively easy...
... Interesting view. I think AJAX works best when you have a RESTfull description of how something holds together. In other words AJAX just helps integrate...
... I would agree - done right. But isn't that always the catch? The hardest thing about REST isn't the architectural style itself, it isn't the URI-space: it...