Let’s say I’m writing an application that retrieves a list of notes from a web service. The URL for retrieving the XML data would most often look something like:

http://whatever.webservice.com/feeds/latest.php?since=timestamp

Which is fine, except one little aspect: it doesn’t tell me if I am getting data back as XML, JSON, CSV or in some other format. I would much rather prefer to have this specified in the URL. Examples of good, clean URLs would be any of the following:

http://whatever.webservice.com/feeds/xml/latest.php?since=timestamp
http://whatever.webservice.com/feeds/latest.php?format=xml&since=timestamp
http://whatever.webservice.com/feeds/latest.xml?since=timestamp

I personally think the last one is the prettiest. There are reasons why the two first might be “better” from a code and performance perspective though:

• The subdirectory method could handle outputting of the data in that format and only that format, not needing any if-checks.
• The query parameter method could handle a lot of different formats in the same file, and would simply call the correct function to retrieve the data in the correct format.

The last one was what I went for in the end. The problem with that is making Apache understand that this “XML” file is really a PHP file. I solved this easily with this simple  .htaccess file, which forces every XML file within the directory you place it in to be run through PHP:

<Files *.xml>
ForceType application/x-httpd-php
</Files>

Sending the correct headers is also a bonus:

<?php
header('Content-Type: text/xml; charset=utf-8');

Worth the effort? You tell me