Unlike web applications, desktop applications are installed and run on each user's computer. They may still communicate with a central server (where the data may be stored, for example) or they may be wholly self-contained with both the application and its data residing on the user's computer.
Running directly on the computer allows a desktop application to take full advantage of the computer's resources, such as memory or the hard disk, and also enables them to interact with other applications and hardware. If the application is stand-alone, it can be run off-line – without access to a network or the internet – which may be useful if the user is mobile or internet connectivity is poor.
The downside is that the software must be installed on each computer individually, and as updates become available, those updates must be applied. If the data is also held locally on the computer, it is normally the user's responsibility to back the data up to protect it from being lost if the computer crashes or is mislaid.