Firstly, it's important to understand that Movie Masher is not itself a content management system. It's only meant to provide an interface for editing content, leaving the storage and retrieval functions to other systems. Communication between the applet and containing systems is strictly via HTTP requests returning XML formatted or (for assets) binary data.
The term 'content' then, in the context of creating mashes, covers both media assets and the metadata information about those assets and how to arrange them into a mash. Pretty much every type of asset that can be dynamically loaded into an SWF can be loaded by the applet as well.
Learn more in Supported Formats.
The assets are loaded into the applet based on information in the XML metadata, which loads first. There are several ways to load this metadata. Perhaps the most straightforward is to include it as part of the initial configuration. But it's also possible to load it incrementally, based on user interaction with controls in the interface. This can be as simple as creating buttons that load media by type, or more complex keyword searching using a text field.
Learn more in and Media Browsing.
The mash data is generally loaded by the applet as part of the initial configuration. Care must be taken though, to assure that metadata for every media and font reference in the mash is also included in the initial configuration.
Learn more in Mash Loading.
Once the user has changed the mash, the data can be passed back to a CGI on the server for storage and later retrieval. This is accomplished using the CGI control which simply posts the XML data to the script.
Learn more in Mash Saving.