I would like to begin by thanking the organisers of this Conference for giving me the opportunity to present this paper today and to participate in your splendid Conference. I feel most privileged.

Let me admit straight away that I am not an expert in health sciences information. My background is in the management of teaching, learning and research in higher education, and in the application of technology to those processes. Nevertheless, I hope that what I have to say will have a good deal of relevance to your particular interest in health sciences information.

I have been asked to speak about the concept of the virtual library, to say whether I think it is a fantasy or has a chance of becoming reality. In fact I want to challenge very strongly the usefulness, the appropriateness, of the whole concept of the virtual library, the digital library, the electronic library. I want to argue that this is quite the wrong metaphor. I want to argue that we would do much better to abandon that metaphor, and to start thinking instead about the processes of teaching, learning and research, or, to take the example of your speciality, about the delivery of health care, and to consider the information and knowledge management issues which arise from those processes. In other words, we need to start in the real world, not with an inappropriate and essentially meaningless metaphor such as that of the 'virtual library'. Later in my paper I want to talk about some of the generic information and knowledge management issues which need, in my view, to be addressed in order to make best use of the enormous technological possibilities which are now available to us, and then at the very end I want to suggest a better way of conceptualising the future.

Conceptual limitations of the virtual library

Let me begin, therefore, with the concept of the virtual library. And I will give you an example taken from higher education in the United Kingdom to explain why I think this metaphor is wrong.

As you may know, higher education in the United Kingdom, as in many other countries, has expanded very rapidly in recent years. Whereas it was once an elite system, it is now much more of a mass system, with around thirty per cent of those leaving school at eighteen going to university. The profile of the student population has also changed markedly. We have many more part-time students, many more mature students, that is to say, students over the age of twenty-one. We have more students studying at a distance, particularly as a result of the globalisation of higher education. There is increasing integration of study and employment. Conventional libraries have felt the strain of this massive change, and this gave rise to a perception that they could not cope with the new demands placed upon them. University libraries were seen to be failing. So about four years ago a review was established which became known as the Follett Review, named after the chairman of the Review group. I was involved myself with the process. The result was really very conservative. To be sure, it took on board the fashionable rhetoric of the virtual library, and put its faith in the fact that sometime, somehow, electronic solutions would be found to the problems facing Britain's students in gaining access to information. But it treated libraries and their problems in isolation. If we are going to deal with a radically transformed, massified system of higher education, we have to start by thinking about the nature of learning and teaching, about the stages in the learning process, and about the nature of the learning environments we need to construct. Then we can start to think meaningfully about how conventional libraries fit in, and about the role of electronic information resources.

This example has a much wider relevance. The changes affecting higher education in the United Kingdom are found in very many countries throughout the world. Governments everywhere are grappling with the need to provide more and more skilled knowledge workers. It cannot be done conventionally. Given demographic trends, it would be necessary to build one new conventional university every week just to keep pace with the current demand for higher education. Clearly this will not happen. We need to do things differently, and I believe that the appropriate application of technology has an absolutely fundamental role in the development of the learning environments of the future. But the starting point must be the nature and purposes of learning in the future, not the problems of libraries. The concept of the virtual library is too abstract and lacks all context. It is not a useful starting point. It presupposes the answer without adequately formulating the question.

In some ways I believe that the use of the term virtual library reflects the continued strength of a very traditional, and I think dangerous view of academic or research libraries. This is the view that such libraries are primarily about preserving accumulated knowledge or cultural heritage. Many libraries have traditionally seen themselves as being, within their own institutions, semi-autonomous empires. They have often, in my view, failed adequately to engage with the needs of their constituents, or engage with the core purposes of their institutions. They have seen themselves as separate. The concept of the virtual library appears to me to maintain that separation. It is almost as if the virtual library is something you can still see or touch. I despair when people say to me, as they sometimes do, "Have you got a virtual library?" The virtual library, let us be clear, is not a place or an artefact. Instead of discussing virtual libraries, I prefer to answer 'No, we do not have what you call a virtual library. However, we do have many students engaged in a process called learning. We are trying to construct appropriate learning environments for them. Those learning environments have many components, such as learning objectives, learning tasks, methods of assessment and feedback, and learning resources. Those learning resources take many forms: tutors themselves, laboratories, modes of communications, books, journal articles, datasets, documents of all kinds. Some or all these components may be delivered electronically. The important question is whether they work together to create the right kind of learning environment. I am not interested in whether or not anyone can identify something they choose to call a virtual library.

I would imagine that what I take to be a wrong approach is much less common within practising health care institutions than it is in universities. Many of you will be focussed on very practical outcomes, and will be expert in assessing the most appropriate ways of supplying the information which your users require. Elsewhere, however, the use of the concept of the virtual library in a very abstract, ill-considered way, has unfortunate consequences. One of these is the widespread feeling that all things electronic are good, and that conventional technologies such as print have less value. This has led many librarians to invest rather unthinkingly in electronic resources without questioning their real value or the demand for them. In the United Kingdom this is also happening to some extent at a national level, where we have a programme within higher education to create what we call the Distributed National Electronic Resource. We are trying to build a critical mass of electronic materials available to all universities across all the major disciplines. There is a strong medical and health component to this resource, consisting of things such as electronic versions of health journals, medical images, and indexing and abstracting services in the medical field. But I think some of this electronic collection building is an act of faith. We think it is going to be important but we are not really sure how, or how soon.

We are also undertaking this collection building with a very poor knowledge base concerning the information-seeking behaviour of users. We know even less about non-use of electronic resources. As policy makers we need to be much more knowledge-based in our activities. We also need to get a grip on the economics of the electronic information industry. We know virtually nothing about this. Certainly we can see benefits in terms of access, communication and so forth, but we do not really know what it will cost. We are pressing forward in a state of extreme innocence, and in a situation where many of the real costs of electronic information, such as networking, are hidden from the end user. I think the very fashionability of the abstract, ill-considered concept of the virtual library encourages this innocent state in which we hold fondly to the belief that electronic information resources will be either virtually free or at least much cheaper than conventional alternatives. In my view there is no foundation to this belief.

First, the abstract concept of the virtual library is not a useful starting point for thinking about real life problems. The virtual library is not a place or an artefact, and it is emphatically not an end in itself. There is no merit in constructing virtual libraries for the sake of doing so. The uncritical use of the term virtual library encourages an equally uncritical attitude to the returns on investment in electronic resources, and nurtures the innocent belief that electronic information is virtually free.

Let me now stress that I believe firmly that the intelligent application of technology to the provision and management of information, of knowledge, will be profoundly important, whether we are talking about education, or research, or medicine, or any aspect of the economy. I am not trying to argue against an electronic future. So what I want to discuss now are some of the barriers to the widespread use of electronic information resources. I shall concentrate on the organisation of access to information, and discuss what I see as a very promising way forward.

Barriers to the use of electronic information resources

Before considering the problems with organisation of access to information let me briefly consider some other barriers.

First, of course, there is physical access, and although I imagine that many of your institutions and organisations are well endowed in terms of hardware and networks, the challenge of maintaining an up-to-date infrastructure is large and costly. This applies especially in universities where students are concerned. In the United Kingdom, the Report of the National Committee of Inquiry into Higher Education which has just been published (the Dearing Report), calls for all students to be equipped with their own laptop PCs by the year 2005. This is immensely challenging, not just because the current rate of ownership of PCs among UK students is only around twenty-five per cent, but also because of the networking and support implications. The challenge of providing adequate access to networks is even greater if we take a global perspective. I learnt recently that there are more telephones in Manhattan than in the whole of sub-Saharan Africa. And as Keith Bezanson has just reminded us, one half of humanity has never used a telephone.

Some of the other key issues, leaving aside for the moment the question of the organisation of access to information, are, to my mind:

authentication;
copyright;
quality assurance
preservation.

So far as authentication is concerned, there is an urgent need to develop national and international authentication services which can verify the identity of a user, verify the service to which that user has access, and provide seamless access to that verified user over a wide variety of authorised systems and services. In UK higher education we are experimenting with services of this kind, but my own feeling is that long-term solutions will come from major service industries such as the banks and the credit card companies.

Copyright remains a major barrier, although I am more hopeful than I was perhaps eighteen months ago that the major publishers are starting to get their acts together, and are seriously addressing issues to do with the ownership and control of information in an electronic environment. I do not believe that an attitude of hostility towards the publishing industry will bear much fruit. It will be a slow learning process on all sides, with small practical steps taking us forward.

Quality assurance remains an important issue. As has often been observed, the Internet makes anything available but it also makes everything available, and in the absence of a proven model for quality-assured academic publishing in an electronic environment, equivalent in certainty to the models which apply to printed publications, there is a sense in which everything that appears on the Web has equal status. This leads to the danger that scholars will retreat into a closed world. The good news is that scholars can collaborate on a global basis. The bad news, in the absence of an established electronic publishing paradigm, is that leading scholars may feel no obligation to communicate with anyone else.

The final point I want to make here about challenges to the electronic future concerns preservation. I do not think we have begun to appreciate how worrying an issue this is. Is it true that eighty per cent of the contents of our research libraries consist of materials printed on acid paper that will ultimately disintegrate? Is it true that seventy-five per cent of the transactions of the American Government are conducted electronically and are not systematically conserved? At the point in history, where our opportunities seem unlimited, are we in fact presiding over the disintegration both of our printed heritage and the electronic heritage of future generations? In all our rhetoric about the information society, we focus on future capabilities, on new forms of producing and distributing information. It may well be that in fact it is the preservation of our cultural heritage, in both printed and electronic forms, which should be at the top of our agenda.

I have made some brief remarks here about some of the barriers to the more widespread use of electronic information resources, stressing issues to do with access, authentication, copyright, quality assurance and preservation. There is a large agenda for us to address. But I would like to focus in more detail on the question of the organisation of access to information services.

A simplified view of networked information systems might consider three aspects. The first is 'conduit' or the network infrastructure itself. The second is 'content', the services that are delivered. The third, and the least well understood, is the concept of 'interconnectedness'. It is this 'interconnectedness' that I want to focus on.

As a very ordinary, mundane example, let us take a group of students who want to discover journal articles and books about, say, the history of Costa Rica. In a good library, they can look on the shelves. They can search the catalogue. They may have access to some databases on the Internet. They can look in databases on CD-ROM. But each of these is delivered through a separate user interface, they may have to move from one machine to another, and they will probably have to write down or print out the results.

Once they have discovered some relevant materials, they will have to find out where they are, or locate them. They may well have to return to the library catalogue and repeat the searches. Or they may look for the materials in neighbouring libraries, in which case they will probably have to repeat the searches yet again.

If they cannot readily locate some of the materials, they may come to the inter-library loan desk and make requests. Probably they will have to write down the bibliographic details again. Inter-library loan staff may then repeat some of the searches themselves as a check. Requests may then be sent for materials, involving more re-keying of the same data.

What we have is a series of horizontally interconnected functions - search, locate, request, deliver - which are not interconnected by systems. Data does not flow across the functional links. The connections are all made by wasteful human effort, erecting barriers to full use. The importance of better integration is becoming clearer as the use of electronic resources increases. The example I have just given refers only to books and journal articles. However, users are likely to have access to a much wider range of resources and services, making the problem of integration that much harder. We need to move to a common, managed framework with a high enough level of interconnectedness to reduce the overhead on users and library staff alike.

The definition of such a framework has been the objective of a project with which I have been associated in the UK. It is known as MODELS, which stands for MOving to Distributed Environments for Library Services. The outcome has been an information architecture which I want to introduce very briefly now.

Consider the three pictures shown in Figure 1. Each picture has three layers. At the top is a presentation layer, where the user accesses resources. In the middle is an organisational layer, a layer which brings services and resources together, and at the bottom are the services and resources themselves. The first picture shows the situation until recently - multiple resources accessed by multiple systems and multiple user interfaces. The organisation layer in the middle is almost non-existent.

The second picture shows an advance over the first, with a web browser providing a more unified presentation layer. However, the organisation of the multiple resources is still very weak, consisting only of a set of links from the same web page. The third picture shows a future situation where the organisation layer, shown by the box in the middle, is fully developed. This intermediate layer we call a broker, and the broker inhabits what we think of as a trading place.

As shown in Figure 2, the trading place provides vertical integration. It is a meeting point for service requests and service providers.

However the broker also provides horizontal integration within the trading place, linking together the functions of discover, locate, request, deliver - as well as authentication and charging - to present an integrated seamless service to the client. This is shown in Figure 3. The broker is a layer of software, or middleware, which provides to the client an organised view of resources, and which is the channel for the interconnectedness, the horizontal flows of information, which I stressed earlier.

The Information Architecture which we have developed shows in some detail the components of a broker, and Figure 4 gives an idea of some of these components. At the top are means of interfacing with the client. In the middle, are means of describing services and mapping requests to and from them, and at the bottom are some of the core pieces of software which manage access to distributed services. You notice that Z39.50 is a key protocol in all this.

This Information Architecture could itself be the subject of a series of papers and presentations, and it is admittedly quite a complex subject. My purpose today is not to describe but to indicate the nature of the problem it is addressing. It is all about integration, interconnectedness, about providing the 'glue' which will hold together the diverse system and services which the user will want to access. And remember that these services are not predetermined. We need broker services which are flexible enough to incorporate new services and eliminate others. So everything has to be built around emerging standards and protocols such as Z39.50.

I believe this model is an extremely powerful tool, and it is now being used in a variety of projects which are building systems of the type I have described. If we can make a decisive move towards a higher level of vertical and horizontal integration of diverse services, then I believe we will decisively extend the use and usefulness of electronic information services.

Before I come to the last section of my paper, let me recall the ground I have covered so far. I have criticised the unthinking use of the concept of the virtual library, and argued that our starting point must be the core business processes with which our institutions and organisations are engaged. Information needs arise from these basic business processes. We can then look at how best to address those needs. I have looked at some of the barriers to the more extensive use of electronic information resources, focussing on physical access, authentication, copyright, quality assurance and preservation.

I have argued that we need a more integrated approach to the delivery of networked information services, and I have presented the model of a broker service as a means of achieving this. Whether or not you like the term virtual library, we need that integration to make a reality of our dreams.

Information landscapes

Let me finally try to bring all this together. At the top of the picture in Figure 4, you will see something called a landscape. I think this is an immensely important idea, and one that is gaining currency. A landscape is the user's personal view of the information universe. It is much more than an interface. It is a kind of map. But more dynamic than a map. It not only shows the user where appropriate and relevant information resources can be found, it also amends the map as new resources are discovered, and dynamically alerts the user to the availability of new things. And it is about more than information resources as we understand them. A landscape also includes modes of communication, such as e-mail conferencing, audio and video conferencing. It is intended to represent a complete working environment, and in that sense it may include the tasks which people need to carry out.

The software to develop this kind of landscape, this kind of working environment, is starting to become available in the context of the Internet. There is already a parallel in the use of the landscape metaphor by companies such as GeoCities, which provide Internet users with facilities to construct their own personal Web sites, or 'homesteads', on the Internet 'frontier'. You can even choose to build your Web 'home' in a specific neighbourhood of like-minded people. The concept of a landscape is evidently a powerful way of enabling people to make sense of virtual territory.

But I also like the landscape metaphor because I see it as being individualised to people's real working needs. Information resources, which are a part but only a part of the landscape, are firmly integrated into people's working and learning environments. They are not separated out, not made to standalone. The concept of the virtual library detaches information resources from real activities. The concept of an information landscape reintegrates them. So my dream is not of the virtual library, but perhaps of something closer to virtual reality - the managed information, communication and knowledge landscape.

Thank you very much for your attention. I wish you will for the rest of your Conference.