Question

“Clean URLs” also known as “RESTful URLs” are user-friendly, purely structural, and do not contain a query string. Instead they contain only the path of the resource.

ie: “http://twitter.com/users/show/”+username+”.json”

Questions about server-side functionality:

1. Do I need to make a unique server-side API script for each directory?

2. Can I forward all requests to just one script if so how do I
   pull useful information from the Clean URL structure ($_GET[‘url_structure’])?

3. Why does twitter call to a .json file
   that surely doesn’t exist. It must get generated on request. How
   does this work? This leads me to believe that the answer to question
   2 is yes.

Answer 1

1) Not if you use a RESTful framework like RecessPHP or if you use a mod_rewrite rule in your .htaccess file to redirect all API requests to a single PHP file (known as the front controller).

.htaccess

RewriteEngine On
RewriteRule ^/api/ api.php

api.php

$request = $_SERVER['REQUEST_URI'];  //this would be /users/show/abc.json

---

2) You can use the rewrite module of apache to redirect all api requests to a special PHP file that handles them. Depending on your apache configuration, the original requested (RESTful) url will be stored in a server variable in PHP, I believe it’s $_SERVER['REQUEST_URI']. Of course you could also just pass along a $_GET[] variable to PHP that contained the RESTful url.

.htaccess

RewriteEngine On
RewriteRule ^/api/([^\.]+).(xml|json|atom) api.php?url=$1&type=$2

api.php

$request_parts = explode('/', $_GET['url']); // array('users', 'show', 'abc')
$file_type     = $_GET['type'];

$output = get_data_from_db(); //Do your processing here
                              //You can outsource to other files via an include/require

//Output based on request
switch($file_type) {
    case 'json':
        echo json_encode($output);
        break;
    case 'xml':
        echo xml_encode($output); //This isn't a real function, but you can make one
        break;
    default:
        echo $output;
}

---

3) Twitter (and many other APIs) use this because it is a convenient way of supplying the format that an application expects back from an API. All of the API requests are rerouted to a single PHP file which handles creating all the files and echoing their contents to the output. The file is never actually stored on the server (unless it is cached).

---

Good Resources

• mod_rewrite, a beginner’s guide
• Create a REST API with PHP
• Recess PHP Framework | Tutorials

---

A note on RecessPHP. It’s certainly a great tool and I would encourage you look at it (maybe at its source to get an idea of how it processes things), but that said, it seems a bit clunky to me. The fact that path names are written in special comments seems very not-PHP to me. I’d stray away from this, and I wouldn’t call it the perfect framework, but it’s certainly a start. Good luck!