What is Hypermedia?

(taken from Chapter 2 of Hypermedia and the Web by David Lowe and Wendy Hall)

We begin this chapter by considering the question "what is hypermedia?" In order to understand how to approach hypermedia development, we need to understand what we mean by the term hypermedia development. In order to understand hypermedia development, we need to understand what we mean by the hypermedia application which we are developing. In other words, by developing an understanding of hypermedia which extends beyond a mere collection of technical wizardry and flashy displays, we will provide a context in terms of our interpretation of the purpose/goals of hypermedia which allows us to develop an understanding of how to most effectively approach the development of the next generation of truly effective hypermedia applications.

The term "hypermedia" is becoming more and more widespread. But what exactly is hypermedia? This sounds like a relatively simple question. It is not! Along with terms such as hypertext and multimedia it is bandied about the press as though it were a cure-all for all information management woes. Despite this, or possibly partly because of it, the term is still rather ill-defined. Before trying to identify or develop a useful definition let us divert for a moment and consider the way that human memory works.

(a)

(b)

Figure -1: (a) Process of writing and reading using traditional linear media (b) Process of writing and reading using non-linear hypermedia.

Hypermedia and Human Memory

Human memory is associative. We associate pieces of information with other information and create complex knowledge structures. We often remember information via association. That is a person starts with an idea which reminds them of a related idea or a concept which triggers another idea. The order in which a human associates an idea with another idea depends on the context under which the person wants information. That is a person can start with a common idea and can end up associating it to completely different sequences of ideas on different occasions.

When writing, an author converts his knowledge which exists as a complex knowledge structure into an external representation. Physical media such as printed material and video tapes only allow us to represent information in an essentially linear manner. Thus the author has to go through a linearisation process to convert his knowledge to a linear representation. This is not natural. So the author will provide additional information, such as a table of contents and an index, to help the reader understand the overall information organisation.

The reading process can be seen as a transformation of external information into an internal knowledge representation combined with integration into existing knowledge structures. These processes are shown in Figure 2-1a. For this the reader breaks the information into smaller chunks and rearranges these based on the readers information requirement. We rarely read a text book or a scientific paper from start to end. We tend to browse through the information and then follow the information trails that are interesting to us [45].

Hypermedia, using computer supported links, allows us to partially mimic writing and reading processes as they take place inside our brain. We can create non linear information structures by associating chunks of information in different ways using links. Further we can use a combination of media consisting of text, images, video, sound and animation to enrich the representation of information. It is not necessary for an author to go through a linearisation process of the author’s knowledge when writing. Also the reader can have access to some of the information structures the author had when writing the information. This will help the readers to create their own representation of knowledge and to integrate it into existing knowledge structures. This process is shown in Figure 2-1b.

In addition to being able to access information through association, hypermedia applications are strengthened by a number of additional aspects. These include an ability to incorporate various media, interactivity, vast data sources, distributed data sources, and powerful search engines. These make hypermedia a very powerful tool to create, store, access and manipulate information.

Definition of Hypermedia

Let us now refine our concept of hypermedia by considering more formal definitions. Table 2-1 provides a number of definitions of both hypertext and hypermedia which provide varying perspectives on hypermedia.

Table -1: Hypertext/Hypermedia definitions

"A memex is a device in which an individual stores all his books, records and communications and which is mechanized so that it may be consulted with exceeding speed and flexibility. It is an enlarged intimate supplement to his memory."

Vannevar Bush, "As we may think", Atlantic Monthly, July 1945

Hypertext: "a combination of natural languages text with the computer’s capacity for branching, or dynamic display"

Ted Nelson, "Getting it out of our system": Information Retrieval: A Critical Review, G. Schechter, ed. Thomson Books, Washington D.C., 1967

Hypertext: "a database that has active cross-references and allows the reader to "jump" to other parts of the database as desired"

Shneiderman, B., Kearsley G., "Hypertext Hands-On!: An Introduction to a New Way of Organizing and Accessing Information", Addison Wesley, 1989

(see also http://www.win.tue.nl/win/cs/is/debra/cursus/definition.html or search the Web for "Hypertext Shneiderman Kearsley" for a discussion of this definition)

Hypermedia: "Multimedia Hypertext. Hypermedia and Hypertext tend to be used loosely in place of each other. Media other than text typically include graphics, sound, and video."

Hypertext: "Text which is not constrained to be linear."

World Wide Web Consortium "Hypertext Terms"

http://www.w3.org/pub/WWW/Terms.html April 1995

"It is tempting to describe the essence of hypertext as its ability to perform high-speed, branching transactions on textual chunks. But this is a little like describing the essence of a great meal by listing its ingredients. Perhaps a better description would focus on hypertext as a computer-based medium for thinking and communication"

Conklin, J. "Hypertext: An Introduction and Survey", IEEE Computer, Sep 1987, pp17-41

"Hypertext denotes an information medium that links verbal and nonverbal information. … Electronic links connect lexias "external" to a work -- say, commentary on it by another author or parallel or contrasting texts -- as well as within it and thereby create text that is experienced as nonlinear, or, more properly, as multilinear or multisequential. Although conventional reading habits apply within each lexia, once one leaves the shadowy bounds of any text unit, new rules and new experience apply."

Landow G, "Hypertext: The Convergence of Contemporary Critical Theory and Technology", The Johns Hopkins University Press, 1996

(see also http://www.stg.brown.edu/projects/hypertext/landow/ht/contents.html or search the Web for "Landow Hypertext")

If we look at these definitions we can find specific points which we might argue with, or can result in confusion. For example, the term "non-linear" is often used when defining hypertext and hypermedia, but can be quite misleading. Any given navigation path through a hypermedia application will be linear - a hard constraint set by the inalterable single dimensional nature of time. It is the explicit support for a network of potential or possible paths through the information which gives us non-linearity, even though a specific actual path will still be linear.

Similarly, many hypertext and hypermedia definitions include references to databases (such as the Shneiderman and Kearsley definition). Again, if taken out of context, this can be misleading. The term database often carries connotations of a high degree of structure, and while this is typical of some hypermedia applications, it is not a defining characteristic (consider, for example, the WWW). A database can be considered to be a repository from which data can be accessed.

Hypermedia is different from ‘traditional’ databases especially from the perspective of the user interaction. Indeed it may be more accurate to refer to hypermedia applications as ‘infobases’, rather than databases. This does not, of course, preclude the use of databases as an enabling technology within hypermedia applications. Indeed the structured information organisation and document management functionality of databases can be effectively used in conjunction with hypermedia style information organisation (i.e. associative linking) to provide a degree of synergy between the two. Database structures can provide a structuring of information spaces which is often overlooked in the traditional hypermedia structures, such as those in HyperCard and the Web.

Indeed, examples of applications, albeit rather contrived, could be found which satisfied most of the given definitions in turn, yet which would not be regarded as hypermedia applications. What then holds up as a definition of hypermedia? The breadth of the ways in which hypermedia is interpreted or defined can be seen in the following two less well recognised quotes:

Hypertext: "is basically the same as regular text - it can be stored, read, searched, or edited - with an important exception: hypertext contains connections within the text to other documents."

Kevin Hughes, "What is hypertext and hypermedia?", May 1994

(see http://www.eit.com/goodies/www.guide/guide.02.html or search the Web for "Web Hypertext Hughes").

Hypertext: "is a web of possibilities, a web of reading experiences. … Hypertext is the language of exploration and discovery"

Charles Deemer, "What is hypertext?", 1994.

(see http://www.teleport.com/~cdeemer/essay.html or search the Web for "Hypertext Deemer").

To provide a basis for the remainder of this book we will answer the question "What is hypermedia?" in an unusual fashion. We will not provide the more usual product based definition, such as most of those in Table 2-1. Rather, we shall provide a goal-based definition (which we will refine later in this book) which allows us to focus on the purpose of hypermedia (and hence the characteristics which we wish to build into applications) rather then the technologies of hypermedia.

To commence this, let us consider the definitions in Table 2-1 anew, not individually, but as a whole. The underlying theme (especially when we consider the intent of the definitions, rather than the content of the definitions) is one of utilising information interrelationships to provide support for flexible access and management of information. For example, why should hypermedia provide support for information interlinking in complex non-linear networks? Surely it is in order to allow us to navigate this information in complex but flexible patterns in order to identify specific items of information or information patterns - the associative linking of information parallels the mechanisms by which our mind retrieves information. Another example: why do we use databases? Surely it is in order to manage the data in a way which facilitates information access and manipulation. Why use multiple forms of media? Surely it is because of the information carrying capacity of the various forms of media.

This leads us to an initial tentative definition of hypermedia as:

Definition: Hypermedia: An application which uses associative relationships amongst information contained within multiple media data for the purpose of facilitating access to, and manipulation of, the information encapsulated by the data.

This definition does not contradict those given in Table 2-1, rather it complements them. This definition is intended to make explicit certain factors relating to hypermedia and its purpose. We will return to this definition throughout this book, refining and utilising it constantly.

Let us then consider this definition. It contains two key elements: access to and manipulation of information; and associative relationships amongst the information. The first of these provides the core tenet which underlies this entire book - the purpose of hypermedia is to provide access to and manipulation of information. This should be foremost in our consideration when we are looking at how to develop hypermedia applications. For the time being we will ignore what exactly we mean by ‘information’ - but we will return to this later, as it is critically important that we understand the idea of information, if we are to develop applications which allow us to handle information.

The second aspect of the definition relates to the associative relationships amongst the information. What makes hypermedia different from other information management systems is that it utilises the complex associative inter-relationships amongst the information which underlies hypermedia applications - hence the common focus on non-linearity and networks. We shall divert for a moment and consider just what these relationships mean in terms of hypermedia.

Hypermedia Linking

Hypermedia systems - indeed information in general - contains various types of relationships between elements of information. Examples of typical relationships include similarity in meaning or context (Vannevar Bush relates to Hypermedia), similarity in logical sequence (Chapter 3 follows Chapter 2) or temporal sequence (Video 4 starts 5 seconds after Video 3), and containment (Chapter 4 contains Section 4.2).

Hypermedia allows these relationships to be instantiated as links which connect the various information elements, so that these links can be used to navigate within the information space. We can develop different taxonomies of links, in order to discuss and analyse how they are best utilised.

One possible taxonomy is based on the mechanics of the links. We can look at the number of sources and destinations for links (single-source single-destination, multiple-source single-destination, etc.) the directionality of links (unidirectional, bidirectional), and the anchoring mechanism (generic links, dynamic links, etc.).

A more useful link taxonomy is based on the type of information relationships being represented. In particular we can divide relationships (and hence links) into those based on the organisation of the information space (structural links) and those related to the content of the information space (associative and referential links). Let us take a brief look at these links in more detail.

Structural Links: The information contained within the hypermedia application is typically organised in some suitable fashion. This organisation is typically represented using structural links. We can group structural links together to create different types of application structures. If we look, for example, at a typical book, then this has both a linear structure (from the beginning of the book linearly to the end of the book) and usually a hierarchical structure (the book contains chapters, the chapters contain sections, the sections contain …). Typically in a hypermedia application we try to create and utilise appropriate structures. Example structures are discussed in more detail in Section 3.2. These structures are important in that they provide a form for the information space, and hence allow the user to development an understanding of the scale of the information space, and their location within this space. This is very important in helping the user navigate within the information space. Structural relationships do not however imply any semantic relationship between linked information. For example, a chapter in a book which follows another is structurally related, but may not contain any directly related information. This is the role of associative links.

Associative Links: An associative link is an instantiation of a semantic relationship between information elements. In other words, completely independently of the specific structure of the information, we have links based on the meaning of different information components. The most common example which most people would be familiar with is cross-referencing within books ("for more information on X refer to Y"). It is these relationships - or rather the links which are a representation of the relationships - which provide the essence of hypermedia, and in many respects can be considered to be the defining characteristic.

Referential Links: A third type of link which is often defined (and is related to associative links) is a referential link. Rather than representing an association between two related concepts, a referential link provides a link between an item of information and an elaboration or explanation of that information. A simple example would be a link from a word to a definition of that word. One simple way of conceptualising the difference between associative and referential links is that the items linked by an associative link can exist independently, but are conceptually related. However the item at one end of a referential link exists because of the existence of the other item.

It is important to note that many hypermedia systems (most noticeably the WWW) do not provide a mechanism for differentiating between the different link types. As a result the same mechanism is used to represent these different types of links. A common result is that users are not readily able to differentiate between the structure of the information space and the content of the information space - resulting in difficulty in navigation.

As a final point on linking, it is useful to note that we could take a holistic perspective and recognise that every item of information is related to every other item of information within some context. Hypermedia however is about sifting through this almost-infinite web of interconnections, to identify and utilise those that help with the goal of information access and manipulation within some context. As was originally recognised by Vannevar Bush [20] over 50 years ago, these associative relationships are analogous to the way in which our mind appears to achieve such a high level of efficacy in information retrieval.

Current State of Hypermedia

Although the underlying ideas of hypermedia are still being actively refined, debated and improved, we are now at a point where many of the technical limitations associated with handling various forms of media, such as image, video, animation, and audio, have been removed - at least to a limited extent. Similarly we have begun to develop an understanding of how to manage the various technologies in a reasonably cohesive fashion (much of this has been a result of the development of the WWW over the last few years). However we have not yet reached an effective understanding of the processes required to develop applications which help us to effectively utilise information. We need to reconsider what hypermedia should be helping us to achieve.

If we look at the current state of hypermedia applications, we can identify at least two major limitations. Firstly, we have yet to satisfactorily address problems of effective location of information. In order to be able to use information we must be able to identify or locate the information we need. Our current applications typically rely on primitive manual authoring of static links, with little subsequent assistance to a user. Similarly current applications rarely attempt to develop an understanding of the users context and how to respond to this.

In some respects this is a core element of the reason for using a hypermedia structure - we identify information based on association. This will however only be effective if the associations are appropriate, comprehensive, and contextual. Although the WWW and most other hypermedia applications support some forms of associative linking, this is limited (it is often neither typed nor structured), incomplete (typically being manually generated, or based on poor automatic indexing) and often lacks mechanisms for providing contexts to aid in comprehension. Additionally, and this is where much recent attention has been focussed, the application should be designed in such a way that assists us in identifying appropriate relationships (i.e. links). This covers factors such as screen and interface design, use of indexes and tables of contents, node or page size, and host of other design characteristics.

Potentially a much more significant problem relates back to the purpose of hypermedia. Most current applications, including the WWW, are to a large extent predicated on Bush’s more restrictive idea of associative information retrieval, rather then Nelson’s broader concept of ‘hypermedia’ being about supporting information usage. This is evidenced by the lack of consideration given to dealing with issues such as information contextualisation, support for intelligent browsing and navigation (beyond mere associative linking in flat networks), information structuring, mechanisms for active annotation and restructuring of networks based on feedback.

The field has evolved in such a way that it has gradually led us towards the current situation of an overly narrow perception of hypermedia as a vehicle for information provision or procurement, rather than the broader concept of a vehicle for information utilisation. The ‘non-linearity’ and ‘interactivity’ which current hypermedia applications incorporate are only implemented in order to help us find an item of information, not to help us use that information. This has been reflected in the development of hypermedia applications, authoring tools, and the entire approach to hypermedia development. This is partially a result of the dominance of the WWW and its basic paradigm of flat arbitrary linking and information downloading. There are, however, positive signs that this is gradually changing, such as discussions on typed links, the use of URN’s (Universal Resource Name) rather than URL’s (Universal Resource Location) and enhanced levels of interactivity and mechanisms for supporting these, such as VRML and Java applets.

In this book we will be considering hypermedia development as the field of developing hypermedia applications which fulfil this wider definition of information utilisation rather than simply information procurement.