The original conception of the electronic library is commonly dated to Vannevar Bush’s 1945 description of the possibilities that technology offered the scientist for the storage, retrieval and manipulation of information (Bush, 1945). His elegant challenge to the “totally inadequate” system for transmitting research remained unanswered for most of the second half of the last century and only in the last few years has there been real evidence that truly electronic libraries are emerging. Much analysis of this evidence implies a radical shift to a new form of information provision. The following (Collier et al., 1993) might be taken as a fair statement of this view:

Such a vision of the electronic library is extreme. The evolution of the electronic library has been, and will continue to be, an incremental process. There is a dividing line between traditional and electronic forms of information service which gradually advances from the traditional towards the electronic. Meanwhile, back in the real world, the modern library works well because it is a hybrid service and will remain so for the foreseeable future.

However, there are a variety of reasons for looking at the electronic library as a self-contained phenomenon. One such reason is tactical. Consider the student at Cornell who, when asked to look something up in the traditional way, replied scathingly, “I don’t do libraries” (Lesk, 1999). For the determined technophile, only the electronic library will do. Clearly, if the modern library is to maintain its central role, it must present itself as a purely electronic phenomenon to such users, while often doing its real work by exploiting traditional library resources under the cover of a virtual façade.

The electronic library is, of course, more than a presentational device. To reach a deeper understanding of the impact of information technology on modern libraries, it is worth considering what terms such as “electronic library”, “digital library” and “hybrid library” really mean. Although these terms are often used apparently interchangeably, quite distinct development strands are emerging. For the sake of argument, it is worth distinguishing an electronic library service such as the California Digital Library (CDL) from digital library services such as the University of Michigan Digital Library or Virginia Digital Library. Electronic library services are set apart from existing collections and institutions and are truly virtual creations. The California Digital Library (CDL, 1999), like its Scottish counterpart the Glasgow Digital Library (GDL, 2000), is an entity distinct from the organisations who came together to create it, the “tenth library” to be added to the nine traditional libraries of the University of California (Helfer, 1999). Similarly, the GDL aims to create a “wholly digital resource” to support teaching, learning and research at all levels in the city of Glasgow, thus placing less emphasis on hybrid services. Electronic libraries in this sense aim to exploit the freedom of virtual collaboration to bring together material separated by institutionally distinct physical ownership and location.

By contrast, the digital library concept represented by the Virginia digital libraries is one in which the identity of parent institutions remains distinct. We should distinguish therefore between the Virginia Digital Library Program (Roderick et al., 1997) and the Virginia (University) Digital Library (University of Virginia Digital Library, 1999 and 2000). Here services are based on information owned and housed in Virginia, such as the rich humanities special collections of the University of Virginia (Electronic Text Center, 2000), and the resources of the Library of Virginia (Library of Virginia’s Digital Library Program, 2000) as well as external services to which the digital library is simply a gateway. Such digital libraries often grow from the digitisation of hard copy materials in local collections and, because they present local collections in a variety of formats, they are more obviously a hybrid service.

A variant on this idea of the digital library can be seen in the UK, where the National Preservation Office in co-operation with the Joint Information Systems Committee (JISC) has drawn on previous work from the Arts and Humanities Data Service to define a preservation strategy for the digital heritage (Beagrie and Greenstein, 1998). Here the goal of digitising hard copy originals is combined with archiving purely electronic ephemera, such as the Web sites of political parties maintained only during election periods. This vision synthesises the concepts of the digital and the electronic library as expressed above and has much in common with the digital preservation aims of CURL’s CEDARS project (“CURL exemplars in digital archives”) (Naylor, 1999). More simply, a digital library such as the University of Derby Electronic Library (UDEL, 2000) can be made up of undergraduate reading materials, exam papers and study skills resources. A digital library in this sense bears a strong resemblance to successful electronic reserve/on demand publishing projects (ACORN, 1998; HERON, 2000).

Digital libraries can also vary in their emphasis on externally supplied services and locally held material. The University of Michigan Digital Library (Lougee, 1998; UMDL, 2000) houses earth and space science materials from external commercial suppliers as well as in-house American heritage material (Price-Wilkin, 1997), but is particularly well known for its opac-based gateway listing of external Internet resources. In the UK we have preferred to maintain listing services collaboratively, setting up a national service such as BUBL (Dawson, 1997; BUBL, 2000) with a brief emphasising quality control and selectivity.

The role of the digital library as a tracing or listing service draws strongly on the traditional library pursuit of universal bibliographic control. By creating online catalogues from local hard copy catalogues and networking them beyond the local library, the information profession took a great step forward towards the creation of a distributed information resource traceable electronically, thus enabling the larger community of scholars to achieve something close to the identification of any published item world-wide. However, it has become clear that there is a real need for a similar level of Universal Bibliographic Control to be brought to bear on the uncontrolled world of electronic networked information. OCLC’s CORC project (Senecal, 2000; CORC, 2000) has explored the applicability of cataloguing or meta-data standards such as Dublin core to electronic resources, thus creating an important tool for the establishment of the electronic library. And once the promise of Z39.50 search facilities is delivered by CLUMPS projects (Stubley, 1997) such as CAIRNS (CAIRNS, 2000), enabling diverse networked opac services to be searched in a single pass, the electronic fulfilment of universal bibliographic control will have taken a major step forward.

But knowing what is published is one thing, getting your hands on it is another. As Web-based end user document supply packages become more prevalent, the contribution of electronic service delivery in this area is more obviously hybrid, consisting of an electronic interface sitting over a non-electronic delivery mechanism. Despite the achievements of projects such as ARIEL (JEDDS, 2000), we still await advances in the area of copyright law, or a new model of scholarly communication within the existing copyright legislation, before the true potential of electronic availability of publications is revealed. This should be no less than the free and affordable delivery of files of genuinely significant academic information to the scholar’s desk top.

The way forward on these issues was shown by writers such as Harnad, who exposed the limitations of the bibliographic control model and commercial publishing system in the electronic library. Acknowledging the escalating price levels of scientific research publications, Harnad (1994) has rightly lamented the role of the commercial publisher, arguing that the scholar should turn to electronic publication as a way of subverting the existing structure of academic publication. Why should any writer give their intellectual property to an intermediary whose function is to limit the free circulation of that material by charging for it? Electronic publication opens up a cheaper and more practical way for the academic to share intellectual property with the scholarly community. Harnad has estimated electronic self-publication costs as being 25 per cent of the costs of conventional publication, and pointed to the success of Paul Ginsparg’s Los Alamos HEP preprint archive (Ginsparg, 2000), where scholars themselves have created an electronic repository of their intellectual property. Standards of bibliographic description are less important in this model of the electronic library than possession of the information described in digital collections.

Such ideas have had great impact. Taking a less radical view perhaps, the Scholarly Publishing and Academic Resources Collection (2000) has formed a world-wide alliance of research institutions, libraries and organizations in order to create new partnerships with publishers. From such relationships, a cheaper, technology-based way to disseminate scientific information is envisaged. By working with scholarly society publishers such as the American Chemical Society, the Royal Society of Chemistry and the German Physical Society, a middle way is taken between the subversive model of Harnad and Ginsparg and the present unaffordable commercial model.

Perhaps the most important lesson to be learnt from the writings of Harnad and Ginsparg is that, in contrast to the traditional library world, the pursuit of electronic bibliographic control does not lead to improved availability of publications. Indeed, to achieve levels of bibliographic control that are taken for granted in the traditional library requires the creation of an electronic library in which material is published and archived by the scholarly community. This would enable librarians to eliminate some of the intractable problems of tracing digital information over the internet. For example, Tim Berners-Lee has pronounced it “the duty of a Webmaster to allocate URIs which you will be able to stand by in two years, in 20 years, in 200 years” (Berners-Lee et al., 1998). Given that the current lifespan of a URL is allegedly 75 days, and that the writings of Jakob Nielsen (Nielsen, 1998) and the respected Graphics Visualization & Usability Center user surveys (GVU, 1994-) cite URL instability as a persistent problem for 60 per cent of WWW users, the duties of Webmasters in maintaining URL stability are clearly not being fulfilled.

Logically, therefore, it falls to the custodians of the electronic library to name and trace digital resources. Miller (1996) has argued that the Digital Library community, rather than Webmasters, must identify institutions of long standing that will take the responsibility for resolving institutionalized names into current names for the foreseeable future, and has praised the work of organisations such as OCLC with its commitment to Persistent Uniform Resource Locators (PURLs) (PURL home page, 2000) as complementary to that of the IETF and W3C on Uniform Resource Names URNs (IETF URN Working Group, 2000). But responsibility for maintaining the electronic library cannot be taken without possession of the resources which comprise it, because to solve a problem one must first own it.

The growth of end user access to bibliographic databases in Britain and the USA further illustrates the importance of creating an electronic repository of resources to enhance service provision to users. In the USA, rather than pay to view journal information mounted outside the educational environment, universities decided to buy in entire data collections under time-limited licences and gave networked access free to end users over the campus network. British universities were not in a position to buy in data sets such as the ISI services for a single institution, and thus went into consortium arrangements such as the CHEST agreement, which created the UK’s ground-breaking BIDS-ISI service. This proved the first in a line of successful hostings of bibliographic databases at a variety of data centres on the JANET network, massively increasing levels of end user access to information (East and Tilson, 1993; East, 1994).

Subsequent developments in UK network management further emphasised this trend. For example, in the late 1990s, the cost of supporting British University Internet access to the USA brought home the need for caching the core of genuinely valuable American information at UK sites (Law, 1997). Against this background arose the concept of a “distributed national electronic collection” (Law, 1995), in which a core of quality electronic materials is defined centrally but maintained over a distributed network of servers. The creation of such a resource should enable the user to exploit a controlled electronic library of valuable resources without trawling through the wastelands of the entire Internet. This idea has recently been enshrined as a cornerstone of national information policy by the JISC, in its vision of a Distributed National Electronic Resource (DNER) (JISC, 1999; Wise, 1999).

This marks a high point in the development of the national electronic library service in the UK. However, crucial questions now remain to be asked about the immediate development of the DNER. Although national consortium agreements have been the driving force behind the acquisition of content for the DNER, such national consortium agreements may not be the best way of acquiring content in future. Recent attempts to reach national consortium deals with commercial publishers such as Elsevier (Elsevier Science Direct, 2000) have been negotiated by brokers working through the National Electronic Site Licensing Initiative (Woodward, 1999; NESLI, 2000). These have been tortuously complex negotiations yielding often unattractive prices for complete portfolios of commercial electronic journals, significant portions of which are of no interest to cash-strapped purchasing libraries.

A better way forward may well be to re-consider attempts at national Higher Education consortium negotiations while pursuing regional cross-sectoral deals in the way that the University of Michigan Digital Library has, by opening up its digital resources to all educational bodies within the local metropolitan area. Such regional consortia are able to bring in finance from Local Authorities, Further and Secondary Education sources and other regional bodies. When this buying power is added to that of the Higher Education sectors, the total income available to commercial publishers from a UK-wide range of regional consortia would be greater than from a monolithic block of UK HE institutions. Such an approach also fits well with the government’s lifelong learning agenda.

If the electronic library is to be a resource shared via cross-sectoral alliances, then the challenge of training an expanded user group has to be addressed. This is not a new issue: a striking feature of the growth of electronic library services is that the role of information intermediaries in supporting and educating users has grown significantly. The proliferation of digital information services has thrown bewildered users back on to the tender mercies of reference librarians who have created extensive user training programmes in response.

This trend is partly a testament to the professional expertise of reference and other public service librarians, but is also a reflection of how much less usable the electronic library is in comparison to the traditional library. Both users and information professionals have tolerated a certain lack of usability in the electronic library as an acceptable trade-off in return for its other benefits. It is also a matter of concern that if the electronic library is so good, why do you need so many people to show you how to use it?

To buy in highly differentiated datasets, which are heavily branded products with their own interfaces, creates a training nightmare. This is inevitable if electronic library resources are seen in terms of access models, rather than as sets of collections which we, the community of users, can own and shape as we wish. However, as increasing amounts of significant data become available within our own network space, we will increasingly be able to mould the tools and interfaces which make these datasets usable and no longer be dependent on outside providers in this regard. We can see the value of this approach in information systems such as the Ohio Gateway, where the emphasis is less on training the user than making the gateway interface more usable (Tiefel, 1995).

Thus, a prime benefit of creating a distributed national electronic library collection will be the control over the interfaces given by such a collection. Quite simply, we will be able to manage knowledge better than ever before. For example, we will be able to optimise our use of Z39.50 protocols, so that a single search interface will link together separate collections. At present Z39.50 is a standard which displays a remarkable degree of diversity in its implementations – the more this standard is managed “in-house” within the DNER, the better.

We will also be able to do more than treat the distributed electronic collection as a giant, searchable catalogue. Many of the collections available over the national network will be highly differentiated, consisting of special formats (image collections, sound collections, multimedia) or specialised subject matter. Rather than being searched as part of a giant search system via a single “common command language”, they will require the development of their own tailored interfaces in relation to their own well-defined user groups, and will be browsed and interrogated in their own right.

Thus, as we get better at interface design and knowledge management, we will remove some of the need for training. Moreover, users themselves will develop their own transferable skills that they can bring to the electronic library – for example, from the entertainment and retail services of the Internet. Once you have used a Web form to buy a pizza with extra pepperoni and anchovies, you have acquired a genuine level of skills in Boolean logic.

However, the totally self-explanatory library does not exist in either electronic or traditional formats, and the need for training in information management skills will unquestionably continue. If the electronic library is to reach out to new, diverse and dispersed user groups, we will need networked courseware packages to deliver such training at the point of execution of an information-seeking task and at a time when it is needed, rather than in a face to face information skills class at a time determined by timetabling practicalities.

The Resource Discovery Network Virtual Training Suite (RDN, 2000) is already moving us in this direction, offering online training in the resources comprising the DNER. The role of the longer-established Netskills programme (Netskills, 2000) in disseminating the skills needed both to build and use the electronic library is also worthy of note in this regard. In this way, ultimately the distinction between user training and context-specific help screens may become blurred, to the immense benefit of the user. Thus, user education in the first two decades of the twenty-first century could become the victim of “disintermediation” just as mediated online searching was in the last two decades of the twentieth century.

Counter to this trend of making electronic library training a dispersed, on demand activity, there is also an important movement towards centralisation and control of the information skills learning process. Glasgow University, for example, has been particularly well organised in its approach, creating a compulsory, examined and certificated IT Induction Programme for all undergraduates, which contains a certain amount of library-related learning material (University of Glasgow, 2000). This has served as an exemplar for other institutions but also raises larger questions about the issue of validated qualifications in IT skills as distinct from information skills. We in Europe already have a supra-national IT skills qualification, the European Computer Driving Licence (ECDL, 2000). But there may be an argument for having an equivalent qualification in information management skills. Otherwise there is a danger of the topic of information management disappearing, relegated to a subsection of an IT syllabus. It is important to emphasise that the skills of information management existed before the existence of IT or electronic library services and these need to be considered in their own right whenever IT-related training is under discussion.

While preserving our own areas of expertise, we in the information profession should not be separatist about our domain. There is currently in higher education a tremendously beneficial movement towards large integrated learning structures, and this trend will not be bucked. JISC has made its commitment to managed learning environments quite clear (JISC, 1999A), holding up the goal of integrating all information systems around the learner. Administrative and financial systems, information resources, learning support and learning environments, can all be better integrated over technology-based platforms to create “joined up learning” for the student. Whatever our vision of the electronic library is, it must be a vision of that library as an organic part of this managed learning environment.

There is clearly a plurality of views about the nature of the electronic library. This paper has attempted to define some of those differing views, and, against a background of national information policy innovation, has tried to indicate what features we should strive for in a viable electronic library service of the future. In the West of Scotland we have attempted a number of important initiatives, such as the Glasgow Digital Library project (GDL, 2000) and the GAELS project – “Glasgow Allied Electronically with Strathclyde” (GAELS, 2000). The first is a truly virtual library service independent of its co-operating institutions, the second more a digital library project based on existing libraries and existing library services. Our experience shows that the difficulties of providing collaborative, large-scale electronic library services against a background of distinctive institutional identities should not be underestimated.

Librarians should be aware that commercial publishers are capable of providing electronic library services independent of institutional constraints, and are thus free to innovate in ways that libraries, tied down both by their separate institutional histories and identities, and by the need to maintain an existing traditional-cum-hybrid library service, are not. There is a clear danger that the real momentum in electronic library service development and electronic learning will be generated from outside the traditional library and university world, by commercial publishers, rather like the momentum in manufacturing industry after 1945 passing from Britain to Germany and Japan. At the very least we can see that the period of UK academic library leadership in this arena based on pragmatism seems to have passed, and fresh momentum is again coming from the USA where important figures in Computing Science (Arms, 2000), Library Schools (Borgman, 2000) and Foundations (Lesk, 1997) are producing more reflective analysis of the way ahead.

In our view, therefore, it is likely that the truly virtual library, an electronic library created in new network space distinct from existing institutional boundaries, will be the most fertile ground for creating the information service of tomorrow. This type of electronic library gives the freedom to manoeuvre and innovate which our closest competitors – especially those outside the educational sector – already enjoy. Whether this belief is correct or not, only time will tell. But it is a vision of the future which we should pursue with certainty of purpose, since it is a future which others will colonise if we do not.