The content management system used to deliver one version of the paper was developed initially for an interactive statistics package. The decision to use a Content Management System (CMS) was driven by the need for users with little programming experience to be able to produce high quality material of a consistent style, without needing to consider how this is actually implemented. Traditionally websites consisted of a collection of pages, each stored in individual files on the web server and with self-contained formatting. To edit the content of the website these files needed to be edited individually, and the new file replaced onto the web server. When using a content management system, the pages are typically stored in a database, with only minimal layout information. When a page is requested, the request is directed to a script or program on the web server, which locates the page, retrieves the content from the database, and then applies the separately held formatting. To edit pages the publisher interacts with the data in the database records, typically via a web-form, although other database management tools provide equally valid editing mechanisms. Through the use of an embedded HTML editor, even the HTML page code can be hidden from the content provider. All user interaction with the website is via web pages. This makes security easier to manage, as the content providing users only need access to an administrative section of the website, in contrast to ftp or shell access to the server that the more traditional web site publishing methods would usually require them to have in order to add material to their site. The content management system used was built from scratch using PHP, on an Apache web server together with a MySQL database to store the majority of the site content.