The interest in personalisation began with online commerce and the need for one-to-one relationships with customers in the early 1990s. Higher education is rapidly moving towards online delivery and mass education, so students could benefit from more personalised services, hence the recent interest in institutional portals, such as uPortal (1), which can personalise and present information. Within this context, the libraries of the Universitat Oberta de Catalunya and The Open University are embarking on a new programme of work to investigate personalised library environments through their respective projects, PESIC and MyOpenLibr@ry. The personalisation seminar organised jointly by the two institutions provided the opportunity to learn from other researchers and libraries that have implemented MyLibrary examples, developed personalisation and gateway tools and researched the user aspects. This is a truly international topic as reflected in the programme of speakers who came from Greece, Catalonia, New Mexico and the UK.
The seminar was aimed at librarians and technologists and focussed on user-centred models including recommender systems, collaborative filtering, roles-based personalisation, and middleware to support portals and personalisation. Presentations included applications in The Open University’s Knowledge Media Institute and Student Services, and MyLibrary case studies in Los Alamos National Laboratory, Universitat Oberta de Catalunya and The Open University Library.
Georgia Koutrika, researcher in the Department of Informatics and Telecommunications of the University of Athens, set the scene with a presentation entitled A personalised perspective on personalisation (2). She touched on the information avalanche, pointing out that as the amount of available information increases and evolves every day at a high rate, providing accurate, compact, and relevant answers to users' requests becomes harder and harder. Information-filtering techniques help cope with the current "information avalanche", but the ultimate response to the problem may be offering "personalised" information access, i.e., access based on the needs, preferences, and current status of individuals. Despite the popularity and the high promises of personalised access, however, a survey of the field indicates that most current efforts originate in the area of Information Retrieval and are limited to dealing with accessing unstructured documents only without taking structured databases into account. Because of this and several other limitations, the current state of the art is far from offering the kind of personalised access one would really envision. Her presentation included a summary of a representative sample of the works presented at the DELOS workshop on "Personalization and Recommender Systems in Digital Libraries", which was organized by the "DELOS Network of Excellence in Digital Libraries". Finally, Georgia touched upon the University of Athens interests in offering personalised access to structured data through technology incorporated inside database systems.
An interesting presentation by John Paschoud from the London School of Economics, on Portals, portals everywhere, why the Interface-to-Everything is not an Interface-for-Everyone (3), provided much food for thought in the tricky area of authentication in relation to personalisation. John assumed the role of devil’s advocate for the seminar, as he noted that the thinking behind such "interfaces to everything" is at risk of missing the point that the intended users are not just students (or academic staff, or researchers, or administrators), but people; and that most people have lives that include several roles, with several corresponding sets of information needs. It is crucial that we do not confuse the role of a person, with his or her identity. He pointed to the experience of a few US universities, where development of institutional information environments have a (short) lead over most in the UK, but that many students reject the offer of including "personal" information content in University Web portals designed for their use. The fact that many individuals have multiple roles impacts particularly on the processes of authentication (process of confirming the link between a person and his/her electronic identity) and authorisation (the process of linking the electronic identity to a set of resources to which the person has been granted access). Some rights of access derive from roles, and some from individual identities. In this context, John outlined the aim of the JISC-funded ANGEL Project, which is specifying a Roles Namespace suitable for use in UK tertiary education, and a resolution model for use in shared authorisation services which can work behind single or multiple institutional portals, and other ‘presentation layer’ services that meet the requirements of the JISC Information Environment.
Monica Bonett’s (UKOLN) presentation An architecture for personalisation of subject gateways based on web services (4) discussed the experience of the IMesh Toolkit project of using the Simple Object Access Protocol (SOAP) as a basis to provide interoperable tools for subject gateways. After briefly introducing both the application area, (subject gateways), and the SOAP protocol, the presentation explored the reasons for the choice of SOAP as an interface, and discussed how SOAP can be used in the context of subject gateways to build reusable tools to manage and share user profiles. Katie Anstock (5) from Fretwell Downing Informatics spoke of the digital library issues which faced the AGORA project team in the 1990s and how those issues were resolved at a technical level. Many of the requirements identified during the project were adopted in the development of Fretwell Downing's ZPORTAL solution.
The Open University’s Marion Phillips (6), Assistant Director of Student Services, discussed how elements of a customised learner service might be developed using web-based technologies. Using examples from the developing online student support service at the UK Open University, a variety of possible techniques were:
Enrico Motta (7) of the Open University’s Knowledge Media Institute outlined how the semantic web and ontologies (an ontology can be seen as a specification of a shared vocabulary, to enable software agents to co-operate on the web) could be used to support radical new personalisation facilities, based on an understanding of the 'meaning' of web resources, rather than simply on past histories and/or current behaviour. Some examples of how the semantic web and ontologies can be used to provide personalisation services were mentioned to illustrate this principle in two domains: scholarly digital libraries and electronic newsletters.
The keynote speech by Clifford Lynch (8), Executive Director of the Coalition for Networked Information (CNI) on Personalization, Privacy, and Distributed Information Resources was delivered via a videoconference link from the Educause office in Washington D.C. He spoke about how personalisation technologies such as recommender systems have become commonplace and well-accepted in certain environments, most notably electronic commerce settings, but in other environments they have seen little use in part due to organisational cultures that place extremely heavy emphasis on user privacy. Libraries in the USA would be included in this group. An additional constraint in current recommender technology is that it works best in massive, centralized systems; whilst this makes business sense (and indeed can represent a competitive advantage) for the commercial services such as Amazon that have implemented them, it is less clear that this limitation is desirable for consumers, and it certainly is not consistent with the information seeking and use patterns of many communities, where users interact with large constellations of distributed, autonomously managed content resources. He examined some of the prospects for extending personalisation technologies to these distributed settings and for accommodating privacy concerns, and described some developments in distributed authorisation technology which may provide useful infrastructure to support new developments in personalisation.
The final session of the seminar focussed on Personalisation in Digital Library Environments. Mariella Di Giacomo (9) spoke on the Web-based Personal Digital Libraries at the Los Alamos National Laboratory. Mariella presented an overview of MyLibrary@LANL, a service which supports digital library users, either as individuals or as groups, with a personalised Web environment that enhances scientific collaboration. In addition to supporting personalised shared environments, the system incorporates the integration of a recommendation system, a Web link checking mechanism, and tools that extend the functionality of Web browsers. A particularly interesting feature of the MyLibrary service is that it gives users the opportunity to create and populate personalised Web environments called libraries. A personal library holds categories called folders that hold a collection of links. These folders can be shared with other users and a user sharing a library with a group of people is able to define the people who can participate and what their rights should be.
Marta Enrech (10) from the Biblioteca at the Universitat Oberta de Catalunya (UOC) presented the PESIC project, founded by the University’s Internet Interdisciplinary Institute (IN3) and lead by the Virtual Library team of the UOC. She outlined the main objective of PESIC which is to define a personalised information system, from the concept of the analysis of the user's information needs in virtual environments, as well as the analysis of technical requirements for the creation and maintenance of the system. Institutional funding will be made available to develop the personalised environment including “my profile”, personalised services and recommenders. This will involve a needs analysis of a pilot group of researchers based in the IN3, system specification and implementation.
Finally, Anne Ramsden (11), seminar organiser and project manager of the MyOpenLibrary Project at the Open University Library, spoke of the issues from the student perspective, the ever burgeoning increase in information resources and resulting overload, as well as difficulties of access to external proprietary databases and content. The MyOpenLibr@ry Project aims to address these issues and deliver a personalised digital library environment that will support the teaching and learning needs of students and staff. The challenges are the development of a user-centred customisable interface, the need to ensure consistent cataloguing of resources and to include the metadata elements to enable personalisation at the course level. It is equally important that our users have a seamless view of the University and should not have to cope with multiple “mypages”. The Open University is embarking on a institutional portal project to deliver functionality like a corporate portal with channels of information for individuals and user groups – calendars, schedules, courses, student services, library, administration etc. This will provide scope for broader integration of personalised library services across the University - as the portal becomes the primary access point for all library users. That way OU students will be able to find information more easily, carry out more effective research and achieve an all round better learning experience.
Speaker biographies, presentations including the webcast of Clifford Lynch’s keynote talk, and links to projects/resources mentioned during the seminar are available from the Open Library website at http://library.open.ac.uk/aboutus/myolib/seminar.htm .Further information about the MyOpenLibrary Project is available also from http://library.open.ac.uk/aboutus/myolib/introduction.htm . Next year, the Open University Library in conjunction with UOC will be organising another similar seminar in Barcelona.
|Michelle Perrott and Anne Ramsden
The Open University Library
Walton Hall Milton Keynes MK7 6AA