REST (Representational State Transfer – difficult to remember, let’s just call it REST) is one of those things – you never heard of it until a month ago and suddenly it’s everywhere you look.

I’ve been working with web services for a little while now. SOAP seems to be the industry standard, though you hear mentions of XML-RPC.

REST is different. The seminal work for REST is not, strangely, a W3C or IETF document – it’s a dissertation for a PhD in philosophy (philosophy in information and computer science, granted).

All sources insist heavily on the fact that it’s not a set standard, but rather, well, a grouping of ground rules/architectural style:

• XML
• but no RPC (Remote Procedure Calling your objects)
• HTTP (GET/POST/DELETE/PUT for write Read/Update/Delete/Update) is enough, you don’t need an extra layer of protocol, like SOAP – the response contains HTTP status codes
• each ‘resource’ (as i understand it’s the entity you query – not clearly defined) has its own URI

Here’s the REST wiki. REST and Rails are often heard together, and since i’m dabbling in Rails more and more, i might give it a shot.

The simplicity of the ‘architectural style’ is attractive – but i need to read more because some bits are still missing.
What about security ? Is this really going to work for any situation ? Will it make it to a standard ? The SOA ideal is to have something that any application/technology can integrate, so acceptance all over is necessary.

The acronym is certainly nice – and also liable to generate a lot of semi-funny geek puns. I’ll refrain from ending with one.

• Previous Entry: Do Androids Dream of Electric Sheep ?
• Next Entry: Hi, i’m a million lines of code