In its cover story on "Intellectual Capital" on October 3 last year, Fortune magazine argued that a major challenge facing corporations is how to figure out the worth of their collective knowledge and maximise its earning potential. Unused yet expensive-to-protect patents are but one example, and it is becoming increasingly clear that vast wealth lies within reach of those who understand how to assess and manage knowledge.

The wealth of Microsoft resides in what its teams of wizards know. If those wizards quit their jobs and worked for you, you would be a lot less concerned if they left their computers behind than if they suddenly suffered from amnesia.

Now consider the school child who can rederive the formula for determining (say) how fast a particular car needs to go before it flips over when negotiating a tricky bend. Compare that child with the one who has memorised all the formulae but cannot rederive them. The difference is in the knowledge of processes, that rich web of interlinked operations that enables creative "workarounds" in the event of a problem-solving impasse.

Who will get higher scores on exams? Often the "memoriser" will perform better, because many exams favour a kind of regurgitation (and we must not underestimate what the "memoriser" knows, because such a person picks up more than just facts). But in the long run, the "rederiver" will prevail, because of a richness of understanding. That is knowledge. As Alan Kay, father of personal computing, said in 1972: "Let's not measure number of exams passed per year; let's measure number of Sistine Chapel ceilings per lifetime."

É«ºÐÖ±²¥

I could easily list what the memoriser knows (for example, a long list of formulae, vocabulary items, and so on). If that list became impressively long, I could put it on a CD-Rom, or send it down the information highway for perusal by others (yawn, yawn). But in the case of the "rederiver", how on earth could I capture what that person knows? And how could I impart that knowledge to someone else? This is surely one of the great challenges of education theory and practice.

Now, is "knowledge" the answer to what "fits" onto CD-Roms, what "sits" on a file server, or what "travels" down the information highway? Most emphatically not! Knowledge is an emergent property which transcends the fixed-size-and-space concepts of media and information, just as it transcends the notion that you can impart it to students by "filling" them up from the teacher's "vessel".

É«ºÐÖ±²¥

The early years of computer-assisted learning fell into that trap, thinking that "filling from a vessel" was precisely what you could do, and fruitlessly trying to "fill up" the poor pupils. The six-year-long debate about the National Curriculum still has not resulted in a clear statement of what knowledge is.

In contrast, decades of work in developmental psychology, cognitive science and artificial intelligence has led to an inescapable conclusion: knowledge is a dynamic process, a vibrant, living thing, resting on shared assumptions, beliefs, complex perceptions, sophisticated yet sometimes crazy logic, and the ability to go beyond the information given. "Knowledge" is the correct abstraction for describing what people communicate to one another. "Content" is not.

Modern specialists in knowledge capture have developed an impressive array of tools for addressing these problems. We know that it is possible to represent the common-sense rules that lead children to go through a brief phase of saying "I gived" instead of "I gave". We can represent complex chains of reasoning such as the way astronomers discovered the formula describing the motion of the planets, or the way investment bankers hedge their bets in the currency markets. We know that there is a consistent logic present in children who think that 63 36 = 30. We know how this mistake arises, why it persists, how to detect it, and how to fix it, forever.

These examples have in common a rich symbolic representational language describing innermost thoughts of learners in more fine-grained detail than previously possible.

A group of researchers led by Enrico Motta within the OU's Knowledge Systems Group has developed ways to capture knowledge using a range of conceptual modelling tools. These enable a knowledge engineer to sketch informally the relations, properties, and processes involved in the problem solving behaviour of what we call a "domain expert". More important, the informal sketching tools are tied to a formal modelling language that allows a rich encoding of knowledge, sufficient to capture a vast range of problem-solving expertise. Knowledge is captured in a modular way so that relevant portions can be used in different contexts. Such tools are contributing to a global project, developing a vast repository of reusable knowledge.

Roger Schank's group at the Institute for Learning Sciences at Northwestern University in Evanston, Illinois looks at how and why students fail, and what motivates them. Time and time again, his group comes up with a mismatch between the expectations of student and teacher. Students can be motivated by stories which illustrate key points, because the mind is geared to pick up analogies between the story and the real-life situation or problem. The trick is in finding a good story, something at which master teachers are experts. Schank's learning environments are dependent upon the students traversing their own pathways, asking questions, finding analogies, listening to stories.

I have been talking about knowledge, so now it is time to say something about media. An advertisement for "presentation software" (which assists in giving public lectures and demonstrations) in the popular computing magazines reads: "He grew up with MTV and 50 cable channels. Unfortunately, he's in your audience today". The point is that this guy has high expectations based on his lifetime of watching slick videos. But the advert also highlights a crucial factor in education, long understood by the Children's Television Workshop, makers of Sesame Street - attention-grabbing graphics matter a great deal.

We need a rich approach to the message and the medium. The message is not just "content", because a deeper analysis of the knowledge is required. The medium is not just a technical format, such as "video" or "CD-Rom". It is the whole presentational style, the user interface, the accessibility, the interactivity. Just as film evolved from a medium of filmed plays to cinematographic wonders, so are telecommunications, computing, and multimedia facilities evolving into new art forms.

É«ºÐÖ±²¥

At the OU, teams of researchers are experimenting with such art forms. Tom Vincent and his colleagues in the Multimedia Enabling Technologies Group developed a "virtual microscope" to allow disabled students to use a polarising microscope. Superficially, it is a CD-Rom containing thousands of photographs of rock samples at different angles and degrees of magnification and polarisation. It is of such stunning clarity and power that other students are using it, as the facilities go beyond a conventional microscope, giving new insights into the study of earth sciences. Other projects include a dynamic visualisation of changing weather patterns, and CD-Roms with large-print and spoken course materials for visually impaired students. Another group of OU researchers in Paul Bacsich's Electronic Media Research Group has explored ways to deliver full Internet capabilities to students who live in distant lands not even connected by conventional phone lines. His group provides a facility for computer-mediated communications, or groupware, based upon an easy-to-use email and conferencing service, "FirstClass". The basis for this environment is a commercial product from the Toronto company SoftArc, and the crucial thing is that Bacsich's group gives access to users anywhere in the world, who might be connected via phone lines, dedicated satellites, high-speed fibre-optic links, or a host of other methods, and who never need know which method they are using. Because the environment is itself a powerful knowledge enabler for thousands of OU students, and because it has been used for everything from teacher training to advanced masters degree programmes in computing, it fits nicely into the knowledge media paradigm.

My last example comes from a project that my colleagues and I ran last summer. For a dozen students who could not attend a psychology summer school, we set up a range of computing and communications facilities to bring the school to them. We provided Internet video-conferencing for guest lectures from abroad, bulletin boards, email, World-Wide Web access to relevant journal articles, and experiment-control software to design and run memory experiments. With the restrictions of 14,400bps modems, we supplied mobile phones for audio conference calls while computers were connected to other groupware services - and a virtual disco . . .

The groups share a belief that knowledge media represents the future of research and development in life-long learning. We have recently proposed a merger under a joint umbrella, the Knowledge Media Institute. We are not interested in multimedia authoring per se, just as we are not interested in the "content" of a CD-Rom. We want to gain insights into what it means to share knowledge; how knowledge can be captured and conveyed; how to exploit new technologies to satisfy the ever-increasing appetite for communication bandwidth; how to meet the needs of mobile students; what new methods will assist disabled students; how to work in groups with students scattered around the globe; how to harness software in the service of human understanding. Our research agenda is large and exciting. We are as interested in corporate knowledge management and intelligent agents as in the needs of students in the wired society. We want to foster life-long learning, but we need to evolve new media, harness them appropriately, and ensure that they are firmly rooted in a rich representation of knowledge.

To give you a feel for the kind of research projects we hope to undertake, here is a representative sample:

É«ºÐÖ±²¥

*Very large virtual classrooms: this project will look at ways to deliver high-bandwidth "virtual presence" capability to hundreds of students on a single course. The focus is on scalability of tutorial methods and workgroup activities, given effectively limitless computing and communications resources.

*The Mobile Student: many of our students need to study while on coffee breaks, on the road, on the move, and in unusual places. This requires us to push the limits of mobile computing and communications technologies to integrate easy-to-use, easy-to-read, and comfortable resources for students. A group of our most mobile (and demanding) students would test the latest generation of courseware which keeps students in touch with course resources and tuition anywhere in Europe.

*WorldWide BA: we are developing the first complete profile of courses which can be taken by students from anywhere reachable via the Internet, and which lead to a bachelor of arts degree recognised by the OU and others. This activity will require the development of new styles of tuition, course presentation, delivery, and facilities management.

*Spiderman Web HyperCruiser: Internet navigation can be tricky. This project will develop a "space-eye" view that lets users visualise the Web in terms of content, topology, access history, popularity, traffic flow, and a rich mixture of other user-defined attributes.

*Agents on parade: What do software agents do when left to themselves? This project will study the behaviour of unsupervised (but monitored) software agents. What will it take for the agents to discover how (or whether) to disable our monitoring?

*Net surfing for blind people: many disabled users have been at a disadvantage on the Internet. We want to develop a suite of software tools which provide full Web functionality for a large range of disabled users including those who are blind.

*Reusable courseware: snippets of text, animation, video sequences and multimedia documents produced for Course X may be of great value not only to Course Y, but also (in arbitrary combinations) to life-long learners around the globe. This project will look at ways to index, store, retrieve, and configure coherent "knowledge modules" for use by different learners in different contexts.

*KMI top 40: we want to develop a multimedia authoring environment that helps schoolchildren design and implement their own community service or entertainment package. These packages will be released onto the net, and their fate will be decided by the public. The top package each month will be awarded a KMI prize and the designers will be offered a KMI summer internship.

*Multimedia in science: the resources for laboratory work in schools are limited. Pupils need to be introduced to new methods of experimental work and analysis that cannot be achieved through conventional laboratory provision. The institute would provide an opportunity to experience multimedia emulations of laboratory experiments gathered from around the world, and an environment in which new creative ideas can be explored with a view to prototyping these ideas on a collaborative project basis.

*Talking Net Newspaper: new communication facilities offer the potential of distributing newspapers in alternative media in more effective and efficient ways. The project would investigate how a national network could be created for local, regional and national organisations to provide immediate access to news.

KMI is not only about blue skies research. There is a strong emphasis on the needs of our current and future student population. KMI will be committed to a throughput of ideas and developments to course teams at the OU. The OU itself is finding that the nature of everyone's roles is changing. Academics, editors, designers, producers and software developers are increasingly involved in one another's work, and the KMI will explore creative ways of intermingling these evolving roles.

Finally, we are seeking significant external funding to enable the KMI to serve as a focus of activity in other relevant arenas. There has been talk recently of the "education superhighway", and we want partnerships with education establishments and industrial concerns to ensure that we help empower the citizen with tools relevant to everyday needs.

We cannot stress too strongly that just the wiring of homes and the creation of "content" is too simplistic a view of how things ought to proceed. The interlinked processes of accessing, sharing, and creating knowledge require significant research and development. Re-skilling society is a major agenda item for us, and we trust that the next generation of life-long learners will grow in an environment that uses knowledge media to radically improve their lives.

Marc Eisenstadt is professor of artificial intelligence at the OU and director-designate of the proposed Knowledge Media Institute.

É«ºÐÖ±²¥

A version of this article is available on the World-Wide Web at by anonymous ftp (in several formats) from kmi.open.ac.uk in directory /pub/documents, and by email from kmi-info@open.ac.uk