What is hypermedia?
Hypermedia can be thought of as a visual, interactive and non-linear medium for communica-tion, which is based on a human-computer interaction paradigm where the user can browse through a database using point and click interaction techniques. Thus, a graphical user interface and direct manipulation (Shneiderman 1983) is essential to hypermedia.
Imagine having access to a large database which contains informa-tion on a wide range of topics, like an encyclopedia. The database can contain text, pictures, animations, sound, and even video record-ings. The distinctive feature of hyper-media is the ability to browse through the material in a variety of ways. This is accomplish-ed by inserting connections between different parts of the material, linking them together. These links can be followed by the user in a very rapid fashion, using point and click interaction techniques.
Typical applications for hypermedia systems include: information dissemination, interactive encyclopaedias, learning, education, reference databases, interactive presentations, simulations, idea processing, writing tools, personal information management, collaboration tools, games and entertainment, like interactive fiction and adventure games.
Hypermedia is an elaborated form of hypertext. The terms 'hypertext' and 'hypermedia' were coined by Ted Nelson in the early 1970's. He envisioned a system called Xanadu (Nelson 1974), where all the literature of the world would be linked together in a gigantic world-wide distributed database. Annotations and cross-references would make it possible to browse through the literature quickly and easily. The difference between hypertext and hypermedia is that hypermedia, in addition to text, makes use of other forms of representation, like pictures, animations, and sound.
Many other researchers and computer pioneers have contributed to the evolution of hypermedia. Douglas Engelbart has made major contributions to the concepts and the technology underlying hypermedia, such as the invention of the mouse pointing device. In the late 1960's Engelbart and his researchers had a working hypertext system running (Goldberg 1987). Work on graphical user interfaces performed at Xerox PARC during the 1970's (Kay 1978) has contributed significantly to making hypermedia practical.
Networks of nodes and links
In hypermedia terminology, associations between different information items are called links, and the individual items are called nodes. A node can contain one or more links to other nodes, forming a network of nodes and links (see figure 2.1). Here a hypermedia network is also called hypermedia structure.
Figure 2.1: A hypermedia network. Nodes which contain information are related using links. Links and nodes can form a network of arbitrary structure and size.
The user can move around the network in many different ways, randomly, purposefully, and so on. In this work however, we will not discuss navigation and different styles of browsing in any great detail. For a treatise on these issues see for instance McAleese (1989).
The ability to move quickly around the hypermedia network is critical to usability and is what makes browsing practical (Akscyn et alia 1988). It should be noted that the very problem discussed in this thesis is partly caused by this possibility; the easier it is to move quickly from concept to concept, the greater is the risk to loose track of what one is doing and to get lost.
In some systems nodes are called cards. This is especial-ly common when the metaphor of the system is based on a note cards model. Links are often called buttons. A button is an area on the screen on which the user can click to traverse a link.
Ways to present information on the screen
Nodes can be presented on the display screen in a variety of ways (Kahn 1989). One approach is to display nodes in separate, possibly overlapping, windows. NoteCards (Halasz et alia 1987) is an example of a system which uses this method. Another approach is to display one node at a time using the entire screen space. This method is used in, for instance, HyperCard (Apple 1987). In addi-tion to these approaches, new models for how to display nodes will be proposed in chapter 7.
Moreover, the contents of a node might be restricted to what can fit on the physical size of the display screen, or it may contain several screens worth of information. In the latter case scrolling and panning techniques are used to access different parts of the node.
A link usually starts from a particular part of a node, for instance a word or phrase, an object in a picture, or an entire picture. The destination of a link can be a node as a whole, or a specific part of a node, like a word, a sentence, or an object in a picture. HyperCard is an example of a system in which the destination of a link always is an entire node, and Guide (OWL 1988) is an example of a system in which the destination is a specific part of a node. Links are sometimes visualized using graphical attributes like boldface text, rectangles, special symbols, et cetera.
Reading hypermedia
Hypermedia differs from printed media in a number of ways. The major differ-ence is that hypermedia is non-linear in its form. It is possible to access the material in many different ways and jump around between different information items, given that the author has provided the necessary links.
While printed material, for example books, use well known design conventions, like table of contents, index, page numbers et cetera, hypermedia systems often lack such well established design elements. The form of hypermedia systems are still experi-mental. Often each hypermedia author uses her own style and conven-tions. Hypermedia is also more abstract in its nature than printed media. Paper exists in the real world. In the simulated world of hypermedia it is not yet possible to physically touch and feel nodes and links.
Often a distinction is made between the author and the reader of a hypermedia structure. In some cases the user of a hypermedia system is both author and reader at the same time. Idea processing and note collecting, for instance, involves both entering information and reading that information. However, in this thesis we will concern ourselves only with reading previously authored hypermedia. Note that reading a hypermedia structure involves traversing the network, as well as reading the actual content of the nodes. In this study we will refer to reading hypermedia as browsing or navigation.
Browsing and navigation
Browsing and navigation are used frequently in the hypermedia literature. What is the difference between these two concepts? Both browsing and navigation describes the process of moving around the hypermedia network by traversing links.
The navigation process aims at finding some par-ticular information item in the database, efficiently and quickly. It is like a captain on a ship who navigates across the sea in order to reach the destination. Browsing, by contrast, is more aimless. The user can browse around a hyper-media system without any specific question or precise idea about what she looking for. Brows-ing is like shopping for Christmas, we often have no specific idea of what we are looking for, we just know that we want to find a Christmas gift. When we browse through a department store, we may eventually come across a suitable present.
We will think of browsing and navigation as 'physically' moving around in a hypermedia structure. The metaphor of physically moving between nodes and links, like walking around in a city, is a common way to think about navigation in hyper-media networks (Dillon et alia 1990). However, this model might not be suitable for all subjects (Krona 1990). In philosophy, for instance, the material might not be straightforward to organize into a physical structure with clearly related parts. Subtle knowledge and complicated reasoning can make the traditional model of navigation unsuitable.
When navigating a hypermedia network the user must be oriented. She must know where she is and she must have an overall view of the world in order to find her way. If she becomes disoriented she might not be able to reach her goal as quickly as desirable. Instead she has to spend time on getting oriented again. This is like the captain at see, who carefully must keep track of where he is. Otherwise he risks to get totally lost. An ocean is especially difficult to navigate due to the lack of landmarks (other than the sun and the stars). In a hypermedia system we can provide clear landmarks which can help the user back on track if she gets disoriented. This will be further discussed in later chapters. Even though it is possible to browse through a hypermedia network without being oriented, the user is likely to end up frustrated and confused if she does not know where she is. Thus, orientation is desirable for browsing tasks too.
Browsing and navigation should be contrasted to querying in database systems. When making a query in, for instance, a relational database system we have to be able to specify the question in verbal terms. Browsing and navigation does not require that we state a specific question. One of the advantages with hypermedia systems over database systems is the we can use a hypermedia system even though we have only a vague idea about what we are looking for.
When making a query to a database we can be almost certain that we have found all items in the database which contains the keywords we have entered (given that the proper keywords have been used when specifying the database). How-ever, there is a risk that we miss information when navigating through a hyper-media system. We cannot be sure that we have seen all information of relevance until every node in the database have been examined. The pos-sibility to specify search strings can thus be useful as a complement to navigation and browsing. Another advantage with querying is that one does not risk to get lost as easily as during browsing and navigation.
An interesting side effect of browsing and navigation is that one might come across information which would not be found otherwise. This is the same phenomenon as when we look up a word in an encyclopaedia and find additional information which catch our interest. A picture, for instance, might attract our attention to a subject we did not originally have in mind.
Hypermedia tools
A hypermedia system consists of a database and a user interface for browsing and navigating through the database. The database contains information in the form of a model of some domain. We will call the database of a hypermedia system for a hypermedia structure. The most common data model for implement-ing hypermedia structures is a network of nodes and links (see above). The user interface of a hypermedia system often reflects this underlying structure in some way. Many systems use the node-link model as a basis for the user interface.
A hypermedia tool consists of an editor and optionally a high level language for creating hypermedia systems. Almost all tools place restrictions on what they can be used for. Most tools are, for example, restricted to a particular model for how information is presented on the screen. NoteCards (Halasz et alia 1987), Hyper-Card (Apple 1987), and ToolBook (Asymetrix 1990), are examples of tools for creating hypermedia systems.
The problem of disorientation, which is examined in this study, is in part related to the basic properties of the general model for hypermedia. However, specific tools and presentation techniques can introduce additional disorientation pro-blems. The majority of disorientation problems discussed in this thesis are related to specific implementation tools and presentation models. However, the distinction between the general level and the tool level can sometimes hard to define.
Advantages and disadvantages of hypermedia
One of the major advantages of hypermedia is the ability to quickly follow associations and look up related material. References can be traced both back-wards and forward in a way which can be difficult and time consuming with printed media. In addition, the user can annotate the material and create new references. Infor-mation can also be structured in a variety of ways. Multiple organizations of the same material allow for specialized structures for different user categories. (Conklin 1986, 1987).
Hypermedia has a strong potential for learning applications since learning by exploration (Papert 1980) might be facilitated in a natural way. The student can browse the material and find new information as she explores a subject area. Con-cepts encountered can trigger new ideas, and chains of associations can be followed in a convenient manner (McAleese 1989). However, this is dependent of how the designer has chosen to structure the system. One approach is to have the students create their own hypermedia systems. Another possible advantage of hypermedia for learning applications is that hypermedia systems is usually considered as fun to use. Even though this might be a result of the novelty of the medium, the potential for visual richness and high degree of feedback could be regarded as positive by the users.
However, ease of browsing might increase the risk that the learner skips through the material much to hasty, and thus get a shallow and fragmented conception of the subject. Also, as discussed earlier, the risk of getting disoriented can result in confusion rather than understanding, especially if the user jumps around between different nodes in a more or less random manner.
An additional problem is that using a hypermedia system involves a certain cognitive overhead (Conklin 1986, 1987). The problems is that the user has to interact with the system in order to accomplish anything, which can be more or less complex. The author of a research paper, for example, might suddenly want to make an note on a new idea which she comes to think of. If this is compli-cated and requires many steps it is possible to loose track of the idea and partly forget it. There is also a risk the one loses track of what one was writing in the first place.