D-Lib Magazine
October 1998

ISSN 1082-9873

Realizing the Hybrid Library

---

Stephen Pinfield BUILDER University of Birmingham Birmingham, UK [email protected] http://builder.bham.ac.uk	Jonathan Eaton HeadLine London Business School London, UK [email protected] http://www.headline.ac.uk/	Catherine Edwards HyLiFe University of Northumbria at Newcastle Newcastle upon Tyne, UK [email protected] http://www.unn.ac.uk/~xcu2/hylife/
Rosemary Russell Agora UKOLN University of Bath Bath, UK [email protected] http://hosted.ukoln.ac.uk/agora	Astrid Wissenburg MALIBU King's College London London, UK [email protected] http://www.kcl.ac.uk/malibu/	Peter Wynne HyLiFe Manchester Metropolitan University Manchester, UK [email protected] http://www.unn.ac.uk/~xcu2/hylife/

Introduction

In a recent D-Lib Magazine article, Chris Rusbridge, the Director of the UK Electronic Libraries (eLib) Programme, argued for the adoption of the "hybrid library" model of information provision.[1]  This paper, following on from and engaging with his remarks, outlines a number of projects, currently being funded by eLib, which are investigating ways in which the hybrid library can be implemented. It discusses key technological and managerial issues emerging from these projects.

The Hybrid Library

The hybrid library is on the continuum between the conventional and digital library, where electronic and paper-based information sources are used alongside each other. The challenge associated with the management of the hybrid library is to encourage end-user resource discovery and information use, in a variety of formats and from a number of local and remote sources, in a seamlessly integrated way. The hybrid library should be "designed to bring a range of technologies from different sources together in the context of a working library, and also to begin to explore integrated systems and services in both the electronic and print environments."[2]  The hybrid library should not, then, be seen as nothing more than an uneasy transitional phase between the conventional library and digital library but, rather, as a worthwhile model in its own right, which can be usefully developed and improved.

This kind of library has been given other labels. The concept of the "gateway library", for instance, seems to be one which describes a similar idea. Discussing the situation at Harvard, Richard De Gennaro explains, "a gateway library does not replace the book collection with technology. Rather, the gateway, like Janus, looks to the documentary sources of the past, even as it looks in the direction of the electronic sources that will be increasingly available in the future."[3]  In these terms, the gateway library and the hybrid library are the same. They describe the "real world" situation where libraries provide access to a range of different media but also express the ideal of greater integration.

eLib hybrid library projects

The five eLib hybrid library projects began in January 1998 and are all due for completion at various times in 2000. They form part of Phase 3 of the eLib Programme which is investigating issues of digital library implementation and integration.[4]   The projects are approaching hybrid library development in different ways and with a different set of emphases, but they have a number of significant commonalities. They all share the basic aim of achieving greater integration and are working with similar or related technologies. They all also are facing the same broad issues, such as authentication, user profiles, user interface design, database interconnectivity, digitization management, and staff development. In addition, all of the projects are developing and managing partnerships with a number of institutions and organizations, within and outside the UK university education sector.

The projects are presented here as a series of separate stories, followed by brief concluding remarks. The staff of the different projects were invited to briefly summarize their project's approach to the hybrid library and then to write about one or a small number of key issues which they are confronting at this relatively early stage in their project's life. Most have written about two or three related issues. These contributions are not intended to summarize the whole of each project but rather highlight what are for each of them important issues right now. Chris Rusbridge has already provided a preliminary overview of different aspects of the hybrid library; these stories suggest how some elements of that may be realized on the ground.

HyLiFe and MALIBU have given their contributions a user focus. They have discussed the hybrid library in terms of user needs and have both also incorporated observations on service and project management. HeadLine has concentrated entirely on the important service and technical problem of authentication. Agora has looked at the broad issues of interoperability and "information landscapes". BUILDER has discussed several of these issues -- service integration, authentication and digitization -- from an institutional angle.  Together, these contributions highlight many (though by no means all) important aspects of the hybrid library as it develops in university education and beyond.

HyLiFe

Introduction

HyLiFe (Hybrid Library of the Future)[5] seeks to establish, test, evaluate and disseminate a knowledge of operating practices for the hybrid library which for the foreseeable future will comprise a mixture of electronic and print services. The project lays heavy emphasis on the needs of users and on evaluation, at different levels, of the use and impact of the new service now being developed. HyLiFe is a complex consortium involving different types of university institutions spread geographically across the UK, although its relevance is expected to extend to other sectors such as public libraries, industry and health. The diverse nature of the consortium enables parallel hybrid library developments to occur in a range of environments with the particular needs of six user groups as the focus of the project.

By making use of existing, readily-available technologies, the consortium members aim for speedy implementation in three annual iterations, with the intention to move to a Z39.50 (or other) parallel searching facility in year two of the project. More precisely, web-based hybrid library interfaces will be developed for:

• Full-time undergraduate students of Geography. This teaching interface is being developed in the library of the University of Newcastle.

• Franchised College students. Development here is underway at the University of Central Lancashire which delivers higher level franchised courses in Further Education Colleges throughout the North West of England.

• Researchers at the  Centre for Urban and Regional Development Studies (CURDS) at the University of Newcastle. This unit comprises over 30 researchers.

• Remote users of Business Link Services in the South West of the UK, supported by the University of Plymouth.

• Practitioners and students in the field of health, affiliated to the University of Northumbria at Newcastle.

• Distributed learners in the University of the Highlands and Islands Project in the North West of Scotland.

These constituencies represent the characteristics and information needs of both full and part-time students, researchers, remote users and different subject groups. Comprehensive evaluation procedures are in place within the project; these have been formulated by staff in the Department of Information and Library Management at the University of Northumbria at Newcastle.  These have been designed to increase the knowledge-base upon which Library and Information Services are to develop in the future, and to feed directly back into HyLiFe itself. The starting-point of the evaluation process is analysis of user needs in the six constituencies described above at the partner institutions.  Thereafter, evaluation will occur at three levels:

• The process and progress of the project itself with regard to communications, project management, achievement of milestones and identification of problems.

• The performance of the interfaces from the technological and usability points of view.

• The organizational and educational impacts of the new services provided by the project. This level considers the implications for teaching, learning and research activity as well as for the Library and Information Service as an organization.

Interface development is underway in the partner institutions and formal evaluation will begin in the autumn of 1998.  The project intends to disseminate its findings widely.

User issues

Although Library and Information Services have long endeavored to adopt a user-centered approach, the issue has now become even more acute because of the greater heterogeneity of user groups within University education. It is recognized that information needs, IT skills and work-patterns of full-time undergraduate school-leavers may differ radically from those of mature, part-time learners working on a remote site. HyLiFe will increase understanding of these differences by virtue of the inclusion, within its user groups, of many patrons drawn from models other than the traditional, full-time undergraduate University education paradigm. On the other hand, early investigations reveal some commonality of experience between the groups, chiefly:

• Users wanted more copies of key print sources.

• Users were frustrated when resources shown as available on the OPAC could not be located.

• IT and information handling skills are often inadequate.

• Users often had difficulty searching the OPAC.

• There is demand for more support and guidance from Library and Information Services staff.

Users will be consulted directly not only about  technical aspects of the HyLiFe interfaces but also about their usefulness in the learning experience. The need for the close involvement of users has been reinforced by clear evidence that their expressed needs are often at variance with those perceived by Library and Information Services staff.

Management and organizational issues

The HyLiFe partners consider that issues relating to the management of the new service deserve at least as much, if not more, attention as technological issues. Successful hybrid library implementation will involve closer integration among stakeholders such as library staff, computing staff, academic staff, and educational development staff. Case studies in participating institutions will assess to what extent organizational structures are changing in order to enable and support co-operative complementary working in the provision of learning support.

Other management aspects include training and development of service staff in order to achieve an appropriate skills mix, support and instruction for users, team-working and the management of change. Ultimately, these issues must be addressed strategically at the institutional level as the hybrid library promises to become a central element of teaching, learning and research activity. It will be possible to identify barriers to hybrid library development which may be cultural, organizational, strategic or personal, as well as technical and financial.

Project management issues

The management of complex, geographically-spread consortia, as with HyLiFe, presents both a formidable challenge and a learning opportunity in terms of research and of practice. For this reason, HyLiFe is systematically monitoring the process and progress of the project itself through regular progress reports from partner sites and archiving of electronic and print communication.

The project benefited in its early stages from the application of a formal risk analysis, which was undertaken separately by each consortium member but which was co-ordinated by the Project Manager. This exercise conclusively identified that the biggest risks faced by the project are the failure of communications among consortium partners, and the loss of experienced personnel before the completion of the project as a result of their being employed on fixed term contracts. It will be noted that this latter risk may be considered as the most significant Knowledge Management issue which HyLiFe (and, indeed, other projects in the eLib Hybrid Libraries strand) will have to address.

The project will test the effectiveness (or otherwise) of contact through electronic mail, electronic discussion lists and newsletters, print, videoconference, teleconference and face to face meetings at different levels. It will increase understanding of the feasibility, pitfalls and benefits of large-scale collaborative ventures which are based on sharing of resources and expertise.

MALIBU

Introduction

MALIBU (MAnaging the hybrid LIbrary for the Benefit of Users)[6] is focusing on the development of institutional models for the organization and management of hybrid library services. These models will evolve from the implementation of prototype hybrid libraries at the three participating institutions: King's College London (lead site), University of Oxford, and the University of Southampton, each with its own particular organizational model. In addition, the MALIBU models will be tested and evaluated in other institutional environments using a number of testbed institutions. The prototypes will be developed for disciplines in the humanities, but both prototypes and resulting models will be transferable to other disciplines. In doing so, the project will face many technical challenges and will address these by integrating existing tools and technologies for resource discovery and delivery.

Management and organization

The hybrid library is more than an interface offering integrated discovery of resources, whether they are in analog or digital format, whether they are local, national or international. Within the MALIBU project, the hybrid library is seen as an environment with physical and virtual services supporting the professional activities of the users at their work place from the discovery of information to the manipulation and analysis of the delivered resources.

Apart from the technical challenges this "integration" offers, there are many management and organizational issues that need to be faced within institutions moving to such a system. The research and teaching community in UK University education complain about a lack of coherence in the existing support framework, at all levels from institutional to national, as a recent study of the requirements of the arts and humanities sector shows.[7] For universities, a range of key management, organizational, personnel and training issues are involved. There are significant implications for the roles of support staff, which are bound to change even further, and for ways to organize, locate and develop relevant support staff. In addition, structures and procedures are needed to manage the process of change itself from the current structure of library service provision to the new hybrid environment.

The MALIBU project will throw light on whether and how organizational structures are impacted by hybrid library environments and vice-versa. It is likely that the most successful institutions will be those that forge strategic alliances among library services, computing services and academic departments. The experiences within the three partner institutions while they implement prototype hybrid library environments will be comprehensively documented and so contribute to the development of models. At the appropriate stage, testbed institutions will become involved in the extensive testing of the models. This experience will also be carefully documented and lead to the further refinement of the models.

To develop core models that will have very wide if not universal applicability, the emphasis will have to be on identifying those elements that will be common to all university education  institutions -- at least in the UK. It also requires the input into the development of the models from a wide variety of institutions with different organizational structures. The three partner institutions within MALIBU all have different organizational models. At King's College London, the Library and Computing Centre are merged into one single, converged organization; at University of Southampton, there is a looser joint-management structure; and in University of Oxford, the Library and computing services are organizationally separate. When the modules are taken further afield into the testbed institutions, the analysis will cover a wider range of institutional types, including at least one "new", teaching-led university.

Requirements of humanities scholars

By focusing on the humanities disciplines, the MALIBU project directly confronts many of the issues around the hybrid library. In most disciplines, and certainly in the humanities, paper-based and other non-digital resources of information will always be of major importance. It seems very unlikely that sophisticated technology for representing them will ever replace the originals and their value to research. In addition, the effort necessary to digitize all these "legacy materials" would be enormous and impossibly costly. Indeed, the term "legacy materials" as used, for example, by Chris Rusbridge,[8] appears to assume an unavoidable and desired transition to the electronic library, that many humanities scholars will not and cannot agree with. One of the characteristics of the humanities community is the resistance of many academics to the use of electronic resources, in part due to years of under-resourcing in terms of training, support, equipment and digitization of materials. At the same time, however, digital techniques can be employed to provide information about sources of all types to humanities scholars across the globe. Digital artefacts are becoming more important in academic study and have led to significant new research results. The deployment of digital resources in teaching has made previously inaccessible materials and information available to students. By definition, therefore, a humanities university library will have to be hybrid for many decades or longer.

From the very beginning, the MALIBU project has involved academic departments in the project and given humanities scholars a significant role in the shaping of their new hybrid library. User panels participate in the selection of resources, technologies, and services, and contribute to the development of new training and support models. Individual and group experience with the hybrid library as it develops will feed back into the project. In specified discipline areas, including History, English, Film Studies and Music there will be specialized work with groups of users. In addition, the project will study how the hybrid library affects the scholarly work practices of the scholar, an area about which very little is known as yet. All these experiences will be documented and will feed into the development of a "user model" as a necessary complement of the management and organizational models developed by MALIBU.

HeadLine

Introduction

HeadLine (Hybrid Electronic Access and Delivery in the Library Networked Environment)[9] is a 3-year project, undertaken by the London School of Economics, the London Business School, and the University of Hertfordshire. It aims to design and implement a working model of the hybrid library in actual academic environments, providing access to a wide range of library materials, regardless of physical form, from a single web interface. The subject focus is economics and business studies although the project aims to demonstrate portability to other subject areas and institutions. The model is user-centered and the system will provide a tailored and supportive environment to the user, complete with feedback mechanisms, allowing the user profile to develop over time. Users will belong to groups, according to their interest profiles and access rights. This emphasis on supporting users in group learning is a key issue for HeadLine and one which distinguishes it from other hybrid library projects.

Authentication

The user-centered model employed by HeadLine defines authentication as a predominant issue, since the system must store information about users and their personal preferences, as well as details of resources they have rights to access by virtue of their membership of a given academic community. In this respect, HeadLine seeks to link otherwise disparate sources of data -- which may range from institutional MIS databases to browser bookmarks files -- that serve to define users' attributes and preferences. Authentication therefore plays a key, dual role: firstly, in securing personal data privacy, and secondly, in controlling which resources may be used. A major goal of HeadLine is the promotion of the "single sign-on" approach for access to networked information resources: once authenticated into a HeadLine system, the user should not be required to supply further passwords for subsequent resource access or requests. Moreover, this facility should function consistently, independent of the user's current location. HeadLine therefore engages directly with the complexities of differing schemes of access control for resources available over an open network infrastructure. The creation of a practical, robust, yet flexible authentication model will depend on its simultaneous ability to address users as individuals and to conform to current and future developments in computer access control technology.

Authentication, as defined by Lynch,[10] establishes the identity of a user within a given context; its companion process -- authorization -- controls the levels of permitted access to resources for that user identity. Network operating systems inherently provide this degree of administrative control; however, their locus of influence is confined to the organizations that deploy them. The open network environment of the Internet operates without any such centralized, enclosing authentication scheme; the new era of ubiquitous networked information services has rapidly exposed the limitations of authentication techniques in common use for the past three decades. Problems inevitably arise as users attempt to move between resources where different access restrictions apply. Traditional techniques of access control employing user identifier and password combinations are becoming unsustainable due to their high administrative overhead and practical shortcomings: loss, unauthorized transfer, or theft. The "password proliferation" scenario familiar to librarians, information managers, IT personnel and computer users themselves is symptomatic of this problem.[11]  A number of current, widely deployed access control techniques, such as IP address filtering, rely upon a continuity of identification rather than upon an absolute identification of an individual. However, these often simply displace rather than resolve the underlying issue: IP filtering works well for on-campus use but locks out remote users, for example.

In working to formulate its own authentication model, Project HeadLine will investigate, evaluate and provide critiques of current and proposed authentication methodologies. It will develop this model with reference to a continuum of identified user authentication needs, ranging from personal preferences and attributes to remote access privileges. Fundamental to this model will be the recognition of complex and overlapping relationships that necessarily exist between users, information providers and institutions that license access to network-accessible resources. The viability of the UK ATHENS authentication scheme[12] will form a principal part of HeadLine's explorations, in particular its claimed potential as a national authentication infrastructure able to encompass not only database services hosted by the national data service centers, but also local and even commercial Internet services.

ATHENS and beyond

A logical authentication scheme that spans geographically distributed servers and multiple types of resources, ranging from databases to individual documents and multimedia objects, ATHENS offers a devolved system for creating and managing user accounts.  It originates in the need to provide a streamlined management tool that addresses such issues as fluctuating user populations and increasing use of commercial Internet Service Provider (ISP) accounts by academic staff and students. Its appeal is enhanced by claims that commercial information vendors contracting with CHEST[13] to supply the UK university sector can "plug-in" to ATHENS, thus potentially eliminating the need to maintain separate user accounts. However, ATHENS still essentially relies on a centralized administration model: despite local site input and control over users' accounts, data must still be submitted to a central system for subsequent replication.  If we look to the Internet itself as an alternative, we see a markedly different trend emerging. A recent development is the emergence of Certification Authorities[14] -- organizations that issue "digital certificates" using strong encryption technologies to provide an electronic verification of the possessor's identity. These can be employed in web browsers to provide individual or corporate proof of identity for transactions that involve access to protected or sensitive data. Such credentials issuers represent the pressure of independent commercial initiatives, rather than any centralized administration scheme.

The current authentication scene is characterized by the transition from an older, intrainstitutional model to a newer, inter-organizational environment that requires interaction amongst members of a virtual community comprising a wide range of different entities: users of information services, librarians, IT personnel, information vendors and so on. HeadLine will address authentication issues within this context, with the ultimate aim of developing an authentication standards framework to present to the information community. HeadLine aims to formulate an authentication model that can encompass the need to store users' attributes and protect data privacy whilst remaining sufficiently extensible to incorporate subsequent developments in third-party authentication methodologies. The creation of an authentication model that can accommodate any content provider as well as any individual user will not be the product of technology alone, but, instead, will mandate negotiation with all relevant participants to help define the grounds for standard practice.

BUILDER

Introduction

BUILDER (Birmingham University Integrated Library Development and Electronic Resource)[15] is an institution-based project aiming to develop a working model of the hybrid library within both a teaching and research context. This involves working towards seamlessly integrating access to a wide range of printed and electronic information sources, local and remote, using a web-based interface, and in a way which will be universally applicable.

BUILDER developments are focused on six inter-related modules, which cover various aspects of the hybrid library. These modules are being developed in parallel, in the first instance, for six specific subject areas; history and archaeology, education, business, physics, sport and exercise sciences, and medicine. Work will be concentrated at the University of Birmingham but will be tested and evaluated within four partner institutions: the University of Oxford, University of Wolverhampton, Westhill College of Higher Education, and Birmingham public Central Library. In addition, BUILDER is working with a number of commercial organizations, such as library system suppliers, data providers and publishers.

During its initial planning stage a number of key issues have emerged which are now being addressed. These include service integration, authentication, database (metadata) management, digitization, and organizational and cultural change.

Integration

BUILDER, like much hybrid library development, is web-focused. The web is seen as the best environment to achieve integration of different services. As Lorcan Dempsey et al. observe, the web already provides "a unified presentation layer" through which many different resources can be accessed. "However, the organization of those resources is very shallow, they may be unified at the presentation layer by being linked from the same web page, but little more."[16]  In contrast, seamless integration can only be achieved where services are accessed through a single user-interface and with a single authentication gateway.

If this level of integration is to be achieved, various technologies and processes need to be implemented which can act as "glue" between services. This requires not only an integrated web front-end but also other forms of "glue technology" which can negotiate with services behind the scenes on the user's behalf and return results through a single user interface. In practice, this glue technology may take a number of forms. It may take the form of a standard protocol, such as Z39.50. On the other hand, it may also take the form of bespoke routines (including high level scripts) which may be used to link specific services in particular ways.

Many enabling technologies already exist which can be used to achieve this "deep" integration. However, in some key areas there is currently an absence of tools. There is, perhaps most significantly, an absence of standards-based ways of communicating with many external services. In an ideal world, these would be supplied by the data providers themselves in the form of an Application Programming Interface (API). API's could be designed and released by data providers which would enable institutional services, such as BUILDER, to create a seamless environment for the user. Here a query would be entered, and then sent to a service via an API, it could then be processed and passed back to the user within the user's local interface.

Many service providers are already beginning to move towards this API model. As they do so, it will be essential for them and partner organizations (such as libraries) to work together on establishing standards. Standards in terms of the technology will be crucial. But other non-technological standards will also have to be agreed. For  instance, conventions for displaying data provided by a third-party will need to be agreed. Data suppliers may expect their data to be displayed with some kind of acknowledgement and branding visible to the user. It will also be essential for users to be aware of the sources they are searching in order to assess the quality of the data being returned.

As part of the BUILDER project, it is planned to develop a series of prototypes in this area. By demonstrating the usefulness of the API-driven approach, BUILDER hopes to play a part in effecting general change. In the first instance, it is planned to create simple APIs for the project's own information products (which include databases of metadata and digitized full texts). In addition, it is intended to encourage the development of API's with service-provider partners within the University of Birmingham. Finally, work has also begun to negotiate with external service providers with an aim to developing at least the framework for workable APIs.

Authentication

One important element of providing a seamless environment for users is streamlining authentication procedures. The aim would be to overcome current problems created for users by a multiplicity of login requirements required by services. In the UK, moves are already being made to reduce the size of this problem with the introduction of ATHENS. This aims to provide a single authentication system for government-supported data services, such as BIDS or EDINA.[17]  However, apart from the problems associated with the organization of the ATHENS system itself (some of which are analyzed in the HeadLine contribution), this does not overcome the problem of the wide range of non-government assisted services (although some may choose to introduce ATHENS compliance).

BUILDER is looking at authentication in the first instance from an institutional angle. The project aims to integrate with existing campus-based systems in order to provide authentication. The University of Birmingham uses Novell Directory Services (NDS) to administer its networked services. Development work is underway to develop an interface between BUILDER and the NDS using a standards-based protocol and incorporating Secure Sockets Layers or similar security technology. When the user logs in at the BUILDER web front door, it should be possible to pass user name and password information to the NDS. The system would then generate a response to confirm whether or not the user is a member of the institution, and also provide additional "user metadata", such as their department and status. The possibility of storing further user metadata within the NDS structure is also being investigated. One of the aims of BUILDER is to provide a "wallet" in which a user's rights and user names and passwords are stored. If possible, this means that authentication could be passed on to third-party service providers automatically without manual intervention from the end-user. It may also create the basis of a "user profile" system, where users are presented with a view of the information which relates to their particular interests.

Databases and digitization

As well as providing access to other services, BUILDER is developing a number of services of its own. At the centre of many of these services are the issues of database design and digitization management. In the early stages of the project, these issues are being explored through a series of pilot activities. In order to assist learning and teaching, retrospectively digitized collections of past examination papers and also taught course learning materials (journal articles or book extracts) are being created. In the research area, the complete run of the journal, Midland History, back to its first volume in 1971, is being converted to digital form. Finally, in partnership with the University of Birmingham Press, a new electronic journal is being developed.

All of these developments involve the co-ordination of digitization strategies with database development. Detailed costing models are being developed for the digitization process, investigating the different technical and procedural options. The storage and organization of the digitized material is also to be investigated along with its associated metadata. A database server running SQL has been installed and is being used as the infrastructure around which the digital collections are being built. A migration strategy will also be devised for all of the files.

Organizational and cultural issues

A significant part of the hybrid library projects (and eLib projects in general) is about promoting cultural change, both amongst information services staff and also end-users. Like all projects, BUILDER has a dissemination strategy to help to achieve this and also an evaluation strategy to measure some of the responses of users to developments. As part of this, BUILDER is also implementing an audit exercise aimed at assessing the views of senior policy makers in the University and also Information Services staff about the hybrid information environment.

Agora

Introduction

Agora[18] is developing a hybrid library management system (HLMS) to provide integrated access to heterogeneous discovery, location, request and delivery services. The HLMS is based on the broker model which was developed by the MODELS project.[19] Agora is led by the University of East Anglia, with UKOLN, Fretwell-Downing Informatics, and CERLIM (the Centre for Research in Library and Information Management) as partners. The project also works with several associate groups: service providers, systems developers, and libraries. There are five library associates: Bath Spa University, University of East Anglia, Heriot Watt University, University of Hull, and Manchester Metropolitan University.

A prototype system has been made available to provide a focus for development. It encompasses a range of social science resources, including abstract and indexing services, library catalogues, archive databases and subject gateways. It will demonstrate two scenarios: cross-domain searching for mixed media with authentication; and search through to delivery for monographs and serials. During the first phase of the project, feedback on the Agora prototype will be sought from professional staff at the library associate sites. Their feedback will inform the development of the first "real" Agora HLMS, which is due to be delivered in Summer 1999. The project will seek to combine the two prototype scenarios in real service environments at the library associate sites. An end-to-end service for mixed media from a variety of curatorial traditions has not previously been undertaken.

Agora broker

The Agora broker is based on the MODELS Information Architecture (MIA).[20] This is a conceptual framework for managing distributed services. MIA suggests some principal components of such systems and provides a common vocabulary for both systems developers and information managers (MODELS workshops have shown that MIA is relevant to archives and museum services as well as to libraries).

In implementing MIA, Agora will be exploring some key areas of concern for distributed service development. This section focuses on issues relating to interoperability and the creation of "information landscapes", which will be addressed throughout the lifetime of the project. It is hoped to find some workable and scalable solutions which will be shared with the relevant communities via targeted dissemination activities. Agora is also working closely with other eLib phase 3 projects.

Interoperability

Most libraries today are, in fact, already hybrid libraries -- they own and subscribe to a range of resources and services which are supplied in a variety of formats and media: print monographs and serials, electronic journals, abstract and indexing services on CD-ROM, music CD-ROMs, etc. Many electronic resources are accessed on remote servers. An increasing number of end-users are also accessing these services from outside the "home" institution: users and services are both distributed.

However, there is currently no uniform way of managing and providing integrated access to these hybrid resources. Users are forced to interact with each service individually and waste time in repeating the same steps to search different systems. At the same time, using different interfaces also increases the risk of inefficiencies -- such as failure to discover relevant resources because of unfamiliarity with one service's idiosyncrasies. In a discover/locate/request cycle, the user will also be forced to re-enter the same data when he or she moves from one stage to the next. Agora is using the Z39.50 protocol (as well as ISO ILL) as the "glue" to link services together and provide true integrated access via a single interface.

It is, of course, recognized that Z39.50 does not automatically solve the problems mentioned above; this is one of the reasons why JISC is investing substantial resources in eLib Phase 3 (and particularly in the "clump" projects which are exploring the practicalities of "large scale resource discovery" using Z39.50).[21] There is a range of organizational issues to be addressed. The principal issue is the varying metadata available from different targets. This has several implications for Agora, including the following:

• When cross-searching a number of heterogeneous services, the number of attributes supported by all targets tends to be very small (e.g. both the EU ONE project and the Canadian vCuc project found just one bibliographic attribute in common across project partners).

• Despite the existence of guidelines, institutions adopt varying practices when mapping local data to the Bib-1 attribute set.

• Serials holdings records are catalogued differently in different vendors' systems; this clearly has a major impact for requesting.

• The varying types and levels of detail amongst records returned from heterogeneous services in one result set could be confusing for users (e.g. bibliographic, archival and subject gateway records).

One possible solution is to ensure that all the service providers conform to the MODELS Profile.[22] In practice, this is perhaps unrealistic, given the spread of services currently included and those which will be involved in future. However, the current initiative to develop an international profile[23] could have a significant impact for future hybrid library management. Both vendors and library and information staff need to be convinced of the major advantages conferred by profile conformance. In the meantime, results may sometimes be erratic. It could be argued, however, that the opportunity to cross search (and request) a wide range of resources offers so many new possibilities for researchers, that they may be prepared to tolerate some irritations. However, metadata interoperability is the key to a fully integrated hybrid library service.

The Agora HLMS is likely to also offer access to non-Z39.50 services, through shallow integration within the web interface. Ways of enhancing integration will be considered during the lifetime of the project, at the same time as taking into account developing technologies such as RDF.

Information landscapes

The concept of information landscapes has emerged from discussions at MODELS workshops.[24] (This is not to suggest, however, that the term has not been used in other contexts.) The concept bears different interpretations, but from a user point of view, it means that instead of being offered a mass of undifferentiated resources and services, a personalized view of the hybrid library is presented, based on the individual user's interests. This view, or landscape, may be constructed in different ways, and, initially, it is likely to be fairly simple -- it could be just a set of web links. A more sophisticated service might dynamically match user profiles against service descriptions (to present an up-to-date view of available resources), taking into account subject interests and access privileges.

One of Agora's aims is to explore the construction of information landscapes as part of the hybrid library management system. A layer of software or middleware, which hides underlying differences and which allows the transparent addition of services, will provide the basis for the landscape. However, there are a number of practical and intellectual issues involved. Firstly, it is not yet clear how user profiles will be maintained. For example, how can a service store enough information about users to provide a useful service, at the same time as protecting their privacy?

As indicated, collection level descriptions are needed to support the construction of landscapes. The Z39.50 Explain service will be used for services to describe themselves to the gateway. However, it is recognized that unfortunately most targets have not yet implemented Explain. Therefore, for pragmatic reasons, Agora will maintain an Explain proxy database to store information about underlying services for those targets. This provides a scalable solution.

Authentication is also needed at various stages to control access to services. Agora will be using ATHENS for this purpose. In some environments, access to locate and request services may be limited to staff users only; in others, certain groups of end-users may be allowed access.

The interface presented to the user will change to reflect the services available as the user moves through the landscape. Dynamic configuration means that only search options supported by the underlying databases of the selected services will be presented at the point of searching.

Growing the communities

Although Agora is principally concerned with developing an infrastructure for managing hybrid libraries, the development of understanding and the acquisition of new hybrid library management skills are equally important to producing a working demonstrator. Therefore, raising wider awareness of the issues and opportunities is of fundamental importance to the project. Consequently, the Agora hybrid library management system will be a key tool for demonstrating potential service scenarios to a range of library and information communities during the course of the project.

Conclusion

In many respects, the issues facing these hybrid library projects are those facing all library and information services. The aspects of their projects the writers have chosen to highlight here are now -- and will continue to be -- at the centre of contemporary information provision. The user-oriented approach, emphasized by HyLiFe and MALIBU, is one which should inform all developments in this area. Similarly, their emphasis on the management of organizational change is crucial in this innovative area. The technical area of authentication (discussed by HeadLine and BUILDER) is one which is becoming increasingly important. Interconnectivity and interoperability (dealt with by BUILDER and Agora) are essential if any "seamlessness" is to be achieved in information provision. The role of information professionals in facilitating the creation of navigable information landscapes (highlighted by Agora) is now also central. All of these activities (as well as others in which the projects are engaged) are working towards developing a more navigable and tailorable hybrid information environment for users.

In order to achieve this, as most project contributions have mentioned, it is also essential to be active in "growing the communities". These projects are as much about effecting cultural change as they are about developing new processes and technologies. This applies first of all to information professionals who will need to acquire new skills and new mind sets in their work. It applies also to decision makers within their institutions (both within and outside library and information services) who need to recognize the potential of these new developments in supporting the teaching, learning and research processes. In addition, commercial providers of systems and data need to continue their moves to provide more open, tailorable products.

Finally, information users themselves need to be assisted to develop new approaches to seeking and using information sources. That having been said, it should also be emphasized that users are not interested in authentication, Z39.50, information landscapes, or APIs in themselves. Nor should they be. They are interested in information, not in the packaging in which it is delivered. They are interested in finding out the answers to questions, not in formulating search queries. Hybrid library development must always keep users' interests at its centre if it is to achieve anything worthwhile.

References

1 Chris Rusbridge "Towards the hybrid library" D-Lib Magazine, July/August 1998, URL: http://www.dlib.org/dlib/july98/rusbridge/07rusbridge.html  

2 Ibid.

3 Richard De Gennaro "Foreword" in Lawrence Dowler (ed.) Gateways to knowledge: the role of academic libraries in teaching, learning and research. Cambridge, Mass.: MIT Press, 1997, p. viii.

4 eLib, URL: http://www.ukoln.ac.uk/services/elib  

5 HyLiFe, URL: http://www.unn.ac.uk/~xcu2/hylife/ . Project Manager, Peter Wynne, [email protected]

6 MALIBU, URL: http://www.kcl.ac.uk/malibu/ . Senior Project Manager, Astrid Wissenburg, [email protected]

7 Sarah Porter and Daniel Greenstein (eds). Scholars' information requirements in a digital age. Consultation draft. URL: http://ahds.ac.uk/public/uneeds/un0.html

8 Rusbridge op cit.

9 HeadLine, URL: http://www.headline.ac.uk/ . Project Manager, John Paschoud, [email protected]

10 Clifford A. Lynch "The changing role in a networked information environment",  Library Hi Tech 15, 1-2, 1997, p. 30.

11 J. Leach Recommended security solutions: results of the study into the requirements for authentication, authorisation and privacy in higher education trusted information systems. February 1998. R1002-2; section 1.4.

12 Mike Johnson User authentication and authorisation in the education sector in the UK - January 1998. URL: http://www.athens.ac.uk/info/authentication.html

13 CHEST negotiates for the supply of software, data, information, training materials and other IT related products to the university and other "post 18" education sectors in the UK and Republic of Ireland URL: http://www.chest.ac.uk/

14 See, for example, VeriSign, Inc.,<URL: http://www.verisign.com/ > and company statement "VeriSign, Inc. is the leading provider of digital authentication services and products for electronic commerce and other forms of secure communications." URL: http://www.verisign.com/about/index.html

15 BUILDER, URL: http://builder.bham.ac.uk. Project Leader, Stephen Pinfield, [email protected]

16 Lorcan Dempsey et al. "Managing access to a distributed library resource: report of the fifth MODELS workshop" Program, 32, 3, July 1998, pp. 273-274.

17 BIDS, URL http://www.bids.ac.uk; EDINA, URL http://edina.ed.ac.uk 

18 Agora, URL: http://hosted.ukoln.ac.uk/agora/ . Project Manager, Greg Newton-Ingham, [email protected]

19 Information about MODELS is available at: URL: http://www.ukoln.ac.uk/dlis/models/

20 The MODELS Information Architecture (MIA): URL: http://www.ukoln.ac.uk/dlis/models/models7/

21 Information about clumps is available at: URL: http://www.ukoln.ac.uk/dlis/models/clumps/

22 The MODELS Profile is available at: URL: http://www.ukoln.ac.uk/dlis/models/clumps/technical/zprofile/zprofile.htm

23 The initiative to develop an international profile is being co-ordinated by the National Library of Canada, and involves a range of organizations; the MODELS profile development is informing this work.

24 Rosemary Russell and Lorcan Dempsey "A Distributed National Electronic Resource? MODELS workshop 6 report." The Electronic Library, 16, 4, August 1998, pp. 231-237.

Acknowledgments

Thanks to Chris Rusbridge and John Kirriemuir for commenting on drafts of this article.

© 1998 Stephen Pinfield, Jonathan Eaton, Catherine Edwards, Rosemary Russell, Astrid Wissenburg, Peter Wynne

Top | Magazine
Search | Author Index | Title Index | Monthly Issues
Previous Story | Next Story
Comments | E-mail the Editor

D-Lib Magazine Access Terms and Conditions

hdl:cnri.dlib/october98-pinfield