There have been no fewer than 4 different proposals floated before, but this one combines flexibility, performance, intuitiveness, and minimal architectural changes, and has only a couple of small drawbacks. Let's take a closer look.

He details the functionality - what it is, what's been decided on, and the advantages of the method. The basic approach is a "lazy" one where PHP will only store the request when it comes in, no filtering or anything, until you want to do something with it (i.e. get it via $_GET, $_POST, etc). When the request for data is made, PHP looks at the encoding settings and formats the data accordingly.

Advantages he mentions for this method include removing the guess-work on PHP's part for which encoding the request is and that it removes some of the overhead by not processing the data until it's actually needed.

Their example pulls an image with overlapping characters to be parsed by their PHP script (needing the GD2 extension). They start off by taking the image into a graphic editor and breaking up the characters to create a "pixel library" of the letters (A-Z) and numbers (0-9) that service uses. This is then used by the PHP script to locate what letters/numbers might be present in the image.

They include in upload form to allow users to give it a try as well as a sample pixel library for the full series of letters/numbers and their points.