This chapter gives an overview of the Evolution Calendar architecture. It describes the model/view split of the calendar into a personal calendar server, or PCS, and the GUI clients that appear inside the Evolution shell. Like other base components in Evolution, the calendar separates the data model from the views or clients. This is done so that multiple clients can access the same calendar data in an orderly fashion and without clashes. For example, the user may be running a graphical calendar client. If he then wants to synchronize his calendar with a handheld device, then the corresponding synchronization program (e.g. a conduit for the gnome-pilot package) will also need to access the calendar storage. It is important that both the GUI client and the synchronization program keep a consistent view of the calendar at all times, otherwise one of them will be left in an inconsistent state if the calendar's data changes unexpectedly. Evolution puts the calendar storage in a daemon called the Wombat and completely separates it from clients who wants to access calendar data. This part of the Wombat is called the personal calendar server, or &PCS;. Clients must contact the &PCS; and ask it to open an existing calendar or create a new one. When a calendar component object (e.g. an appointment or to-do item) changes in the &PCS; it will notify all the clients that are using the component's parent calendar.