introduction multimedia
[] readme preface 1 2 3 4 5 6 7 appendix checklist powerpoint resources director

talk show tell print

2.2 hypermedia

Given an information space we may turn it into an information hyperspace, that is, following  [Spaces],

information hyperspace


the logical information space may further be structured in a logical information hyperspace, where the clues become hyperlinks that provide directional information, and the information space can be navigated by the user following directional clues.

In other words,

information is chunked, and each chunk is illustrated or made accessible by an example (hypernode) ...

Now, what exactly does information hyperspace mean? To answer this question, let's briefly look at the history of hypertext and hypermedia.

history


 Bush' (not the presidents') seminal paper As we may think may be regarded as the origin of what is known as hypertext with which, even if you don't know the phrase, every one of you is familiar, since it is (albeit in a rather simple way) realized in the web.

The phrase hypertext was invented by Ted Nelson (not patented, as far as I know), who looked for a less constraining way to organize information then was common in the educational system he grew up with. But before that, Douglas Engelbarth, who incidently invented the mouse, developed the Augment system to boost the human intellect. What for, you may ask. Let me quote the series of flashes that Engelbarth went through, according to Dust or Magic  [Magic]:

classification of hypermedia

 Perhaps it is good to know that Bush (not the president) wrote his article when working for an information agency in the second world war period. From that perspective, we can easily see that hypermedia (combining hypertext and multimedia) were thought of an instruments of intelligence.

Basically, hypermedia systems must be able to deal with:

hypermedia systems

Far from being a definition, this characterization gives some insight in what functionality hypermedia systems must support. Recall that dealing with complex information is what hypermedia is all about.

Is this a natural way to deal with information you may ask. Just think about how you are taught to deal with information and how you actually go about with it. Speaking about Ted Nelson, quoting  [Magic]:

everything is intertwingled


... he realized that this intertwingularity was totally at odds with the education system he spent so long in and had been so uncomfortable with.

Quoting Ted Nelson himself from his book Literary Machines:

A curriculum promotes a false simplification of any subject, cutting the subject's many interconnections and leaving a skeleton of sequence which is only a charicature of its richness and intrinsic fascination.

Judge for yourself. Would you prefer to have an 'immersive' course in multimedia rather than a more or less ordered collection of abstractions?

True enough, the visions of the pioneers of hypermedia where overwhelming. Nevertheless, the concept of hypermedia, that is non-linear media with machine-supported links, or 'text' as a network, found an application in a large variety of systems, see  [Hypertext].

classification of hypermedia systems

  • macro-literary systems -- publishing, reading, criticism
  • problem exploration tools -- authoring, outlining, programming
  • browsing systems -- teaching, references, information
  • general hypermedia technology -- authoring, browsing, collaboration
  • embedded hypermedia -- CASE, decision support, catalogs

 [Hypertext]


An example of a hypermedia system that has extensively been used in education, for example biology and chemistry classes, is the Brown University Intermedia system of which a brief characterization is given below.

Intermedia

  • web = documents + links + maps

retrieval by attributes

  • block -- to apply filters
  • link -- conditional traversal
An interesting aspect of this system is that the user may create maps, that is structures containing documents and links, which form a personalized version of the web of information for a specific user, superimposed on the information space offered by the system.

Dexter Hypertext Reference Model

After many years of developing ideas and exploring implementations, one group of experts in the field came together and developed what is commonly known as the Dexter Hypertext Reference Model, named after the location, actually a pub, where the meetings were held.

Dexter Hypertext Reference Model

  • components, links and anchors
The Dexter model offers an abstract description of hypertext. It made a distinction between components, anchors within components and link between components, attached to anchors. The model was meant as a reference standard against which existing and future hypertext systems could be compared.

Components have the following attributes:

component

  • content -- text, graphics, video, program
  • attributes -- semantic description
  • anchors -- (bi-directional) links to other documents
  • presentation -- display characteristics

compound

  • children -- subcomponents
The Dexter Hypertext Model has been criticised from the beginning. Among others, because compound documents where not adequately dealt with. And also because it did not accomodate multimedia (such as video) content very well. In practice, however, the Dexter model has proven to be even somewhat overambitious in some respects. For example, the web does (currently) not support bi-directional links in a straightforward manner.

Amsterdam Hypermedia Model

When looking for alternatives, a Dutch multimedia research group at CWI proposed to extend the Dexter model with their own multimedia model (CMIF), an extension for which they coined the name Amsterdam Hypermedia Model.

Let's look at the (CMIF) multimedia model first:

(CMIF) multimedia model

  • data block -- atomic component
  • channel -- abstract output device
  • synchronization arc -- specifying timing constraints
  • event -- actual presentation
What strikes as an immediate difference with respect to the hypertext model is the availability of channels, that allow for presenting information simultaneously, and so-called synchronization arcs, that allow the author to specify timing constraints. Also, events are introduced in the model to deal with user interactions.

authoring

  • structure -- sequential and parallel composition
  • channels -- presentation
With respect to authoring, the model supports a declarative approach to specifying sequential and parallel compounds, that is in what order specific things must be presented and what may ocuur simultaneously. Again, channels may be employed to offer a choice in the presentation, for example a dutch or english account of a trip in Amsterdam, dependent on the preferences of the (human) viewer.

The Amsterdam Hypermedia Model (AHM) extends the Dexter Hypertext Reference Model in a rather straigthforward way with channels and synchronization arcs.

Amsterdam Hypermedia Model

  • contents -- data block
  • attributes -- semantic information
  • anchors -- (id, value)
  • presentation -- channel, duration, ...
Obviously, the difference between Dexter and AHM is primarily the more precise definition of presentation characteristics, by introducing channels as in the (CMIF) multimedia model. Another (major) difference lies in the characterization of compounds.

compound

  • children -- (component, start-time )
  • synchronization -- (source, destination)
Each component obtains a start-time, that may result from parallel or sequential composition and synchronisation arcs. Yet another interesting concept introduced by the Amsterdam Hypermedia Model is the notion of context. What happens when you click on a link? Does everything change or are only some parts affected? Then, when you return, does your video fragment start anew or does it take up where you left it? Such and other issues are clarified in the Amsterdam Hypermedia Model, of which we have omitted many details here.

It is perhaps interesting to know that the Amsterdam Hypermedia Model has served as a reference for the SMIL standard discussed in section 3.2. If you want to know more about the Amsterdam Hypermedia Model, you may consult  [Hypermedia] or  [AHM].

research directions -- computational models

Today, hypermedia functionality is to some extent embedded in almost all applications. However, to realize the full potential of hypermedia, and in effect the networked multimedia computer, there are still many (research) issues to be resolved. To get an impression of the issues involved, have a look at the famous seven hypermedia research issues formulated by Halasz.

research issues


  • search and query -- for better access
  • composition -- for imposing structure
  • virtual structures -- on top of existing structures
  • computation -- for flexibility and interaction
  • versioning -- to store modification histories
  • collaborative work -- sharing objects with multiple users
  • extensibility and tailorability -- to adapt to individual preferences
See  [Hypermedia], section 2.3 for a more extensive description. Although the research issues listed above were formulated quite early in the history of hypermedia, as a reflection on the requirements for second-generation hypermedia systems, they remain valid even today. Without going into any detail with respect to the individual research issues, I rather wish to pose the grand encompassing research issue for the networked multimedia computer: What is the proper computational model underlying hypermedia or, more generally, for applications that exploit the networked multimedia computer in its full potential? Some directions that are relevant to this issue will be given in section 3-3 which deals with the multimedia semantic web.
[] readme preface 1 2 3 4 5 6 7 appendix checklist powerpoint resources director
eliens@cs.vu.nl

draft version 1 (16/5/2003)