Hypervideo is the adaptation of the hypertext metaphor to video. By annotating and referencing video objects, diverse media, and pieces of information the video stream can be unlocked to the global web of knowledge. Hypervideo combines the advantages of hypermedia as a dynamic, associative, and extendable network of information that can be shared and searched by many users at the same time, and of video as a natural and intuitive media to convey complex dynamic processes. Therefore, hypervideo will be one of the most important media for learning on the Semantic Web. This article will first introduce the concepts of hypervideo. Then the history of hypervideo applications will be outlined, and current applications on the market will be presented. Design aspects will be discussed taking the example of one hypervideo application. Then, use cases for hypervideo will be presented. Finally, we will discuss possible future trends before coming to the conclusions.
Hypervideo is an application of the hypermedia concept. However, in contrast to other hypermedia, such as the World Wide Web, hypervideo not only is a hyper structure of linked media pieces, but it has an inherent structure. All information linked to the video is organized along the timeline of the continuous video stream guiding the reader and giving a major direction of exploration. To understand the specific nature of hypervideo we should first have a look at the underlying concepts: hypertext, multimedia, and hypermedia (a detailed discussion can be found in Myers, 1998).
Hypertext has been defined as a combination of natural language text with the computer’s capacity for interactive branching, or dynamic display of a non-linear text, which cannot be printed conveniently on a conventional page (Conklin, 1987). The single pieces of information linked together are called the nodes. Anchors are used to mark nodes or sections of nodes and to serve as connection points for the links that actually describe the connection between information entities (Steinmetz, 1995).
Multimedia describes the computer-based integrated generation, manipulation, and interaction of different media, where, according to Steinmetz (1995), continuous (time-dependent) and discrete (time-independent) media have to be combined. Schulmeister (2002) further emphasized the importance of the interactive manipulation of such media as relevant for characterization.
Hypermedia unifies the concepts of hypertext and multimedia by describing a hyper structure of multimedia nodes (e.g., text, image, audio, and video). Common hypertext structures found in the Web combining different pieces of media in a basically text-oriented structure are often referred to as hypermedia (Nielsen, 1990).
The basic model for hypermedia leads back to the early hypermedia systems such as Augment or HyperCard and has been published by Halasz and Schatz (1994) known as the Dexter hypertext reference model. Figure 1 shows the Dexter model that separates the hypermedia structure into the within-component layer (containing the original data units, that is, files that are unaware of their belonging to a hyper structure), that are referenced by anchors. The storage layer represents the hyper structure with the relations between units of information, which then is presented by the run-time system.
The Dexter hypertext reference model
The expression “hypervideo” finally has been coined by Locatis, Charuhas, and Banvard (1990). In a hypervideo the non-linear hypermedia structure is applied to the originally linear video stream, which thus can be broken up. Table 1 summarizes the terminology used to describe the basic components of a hypervideo (Finke, 2005; Finke & Balfanz, 2004).