Most web sites offer a smorgasbord of media types to their users. Typical web content these days includes Portable Document Format (PDF) files, word-processing documents, audio files, movies, and Extensible Markup Language (XML). In some cases, these alternative file types are stored in databases as binary data , or streams of unencoded bytes. Web developers cannot always provide their users with straightforward hyperlinks to these files for downloading. The user chooses a link or enters a URL in the browser's location field, and a servlet or some other web component downloads the binary data to the client. The client in most cases saves the data as a file, for use later in document viewers and other applications.
The following recipes describe how to initiate this download method. In a typical scenario, the servlet sets up a download whereby the browser prompts the user with a "Save As" dialog allowing him to save the files on his own filesystem. These strategies, however, do not guarantee 100-percent consistent behavior among web browsers. Some browsers allow the user to precisely configure how he wants to handle handle certain file types (such as a PDF document). For example, Opera 5 gives the user all kinds of options for dealing with downloads, such as opening up an external helper application, displaying a file using a plug-in, or immediately downloading a file to a specified folder without first opening up a "Save As" window.
The alternative to these strategies is simply providing a link to a PDF file ”for instance ”if the data exists as a file. This method, however, also places a great deal of the responsibility on the client browser, and the user who configures the browser, for managing the media type. While advanced users may like this approach, it is rarely sufficient for less- experienced users.