To retrieve data from a WWW server, a WWW client needs to be able to communicate with it. Typically, a client is a standalone program that relies on some underlying communications package. On a Unix or Windows-NT system this is built in to the basic operating system. On an MS- Windows system or an Apple Macintosh, it isn't built in as standard, and you may need to install the networking software too. There are many excellent guides on how to do this, and your Internet provider may even do it for you. In the MS-Windows case you'll also need to tell the browser where to find the networking software - this is described in section 4.14.
The client is a fairly complex system which not only accesses WWW servers across the network using HTTP, but must understand the MIME and HTML in any responses (and any HTTP error responses too).
Its main task is to display as well as possible the documents you access. Typically, the client application doesn't comprehend all the many different media, text, graphics, audio, video, etc, since that would make it an enormous, monolithic program, possibly unable to run on any but the most powerful workstations. Instead, it is designed to launch the appropriate viewers for the appropriate media types according to a set of configuration information based on the MIME setup for reading multimedia mail.