Second, several studies have proposed high-level frameworks for authoring AR content [12,13]. These studies aim to create an authoring platform that is independent of specific applications or target hardware setups. However, while the studies considered user interaction, most did not deal with the integrated contexts of users and their situations. They confined user interactions to simple contexts such as user input to trigger a transition of the AR content. In the ubiquitous media environment, a smart multimedia service should consider more complex user interactions. Thus, a sophisticated framework is necessary for a user-centric service. Third, the 5W1H-based metadata schema has been utilized in of some studies of context-aware services. The efficiency and usefulness its structure, composition/decomposition, interchanges, unification capability, extensibility, and scalability have been proven for user-centric services [14]. In a service for tourists, one of the most important requirements is ‘context-aware information’ [15,16]. This metadata schema classifies user information from sensors and a user profile into six categories and identifies user situations clearly [14]. It is also easy for everyone to understand and is thus more efficient in terms of management than other metadata structures. Although Alex et al. (2010) holds that “what, where and how” are the basic properties of most AR applications, they did not systematically apply these concepts to their open architecture for authoring and AR content delivery.