As regular readers of 'Ariadne' will know, the fourth annual ELVIRA conference has just taken place at Milton Keynes. The following article is based on my general impressions of the event. A more detailed and complete account can be found in the collected papers, which have been published by Aslib  . The 'extended abstracts' originally submitted for review are online at the ELVIRA Web site .
In the keynote address to the conference, Brian Cook (Griffith University, Queensland, Australia) identified the issues facing people working in the electronic (aka digital/virtual) library field. The list appears daunting: the growing information needs of all levels of society; the impact of the increasing involvement of new players (the entertainment industry, cable and satellite companies) in the provision of electronic information; technological issues (including the need for standards); issues relating to the changing roles of information creators and providers; the lack of consensus in the areas of electronic publishing and copyright. Nevertheless, Professor Cook's presentation was upbeat and positive. Many existing problems could be solved by a collaborative approach. He urged a greater synergy across research projects, and the establishment of consortia (for example, Australian Universities have combined to negotiate a single license agreement with a publisher for an electronic version of one of its journals). Above all, greater attention should be given to users. Researchers should ensure that developments are not purely technology driven, and that they relate closely to real information needs.
The conference provided evidence that there is already movement in this direction. The variety of papers indicated a wide range of research activity - from initial exploratory studies to mature, large scale projects. Technological issues were covered, but also legal and economic issues, and many papers focused on user needs.
A number of papers described the development of new systems where the study of users and their requirements is playing an integral part in their development. The MUMLIB project (Deirdre Ellis-King, Dublin City & County Public Libraries) involves a user centred approach to the design of multimedia resources. This has been applied to the development of the Wordsmiths CD, which profiles a number of contemporary authors from Denmark, Ireland and Portugal. The process began with user trials of similar products. The findings from these informed the development of a prototype: for example, a basic Help facility was created for users who were not familiar with Windows conventions, and browse facilities were developed because it was seen that most users preferred browsing to query-based searching. User trials were also used to evaluate the prototype. This led to several amendments of the system, though the overall assessment was very favourable. A clear, well designed interface and good content were seen as prerequisites to a good system.
This welcome focus on user-centred development was also strongly evident in other projects' presentations, such as those of PATRON  and SuperJournal  (both 'eLib' projects). Christine Baldwin's SuperJournal presentation discussed the findings of a baseline study to examine stakeholders' initial expectations of the system. The prospects of improved access, full-text searching and the use of multimedia were welcomed. At the same time there were some concerns, over (for example) the possible influence of the electronic format on writing style, and the problem of identifying the good quality sources of information from the rest.
Brian Whalley (Queen's University, Belfast) described the setting up of a new electronic journal, Glacial Geology and Geomorphology . Surveys of potential users were carried out to identify how the electronic format could best provide added value. The rapid publication of papers (including colour diagrams) and correspondence were seen as definite advantages. As in the SuperJournal project, there was some uncertainty over the quality of much Web-based material, so publication by an established publisher and (in particular) a system of peer review were seen as important.
Clearly, user studies are also essential to monitor and evaluate services once they are running. Caroline Lloyd (London School of Economics) gave an excellent presentation on one such study, for the DECOMATE project (Delivery of copyright material to end-users). A variety of quantitative data (from interviews and focus groups) and qualitative data (from electronic questionnaires and analysis of log files) was collected and analysed.
David Zeitlyn and Jane Bex (University of Kent) used a micro-study technique which examined key stokes and video recordings of clients using a library OPAC. Breakdowns in the interaction between user and OPAC were examined closely. The analysis called into questions some of the assumptions - made by both system designers and librarians - about the extent to which users actually understand such systems. Zeitlyn and Bex's double-act presentation brought out the need to ground our high technological aspirations in the reality of users' technical competence and training needs.
David Nichols and Martin Twidale (Lancaster University) described an attempt to give library users an opportunity to make contact with other users who were interested in the same materials ("collaborative browsing"). The research was directed at printed materials in the library so that it could be conducted without the necessity of amending library system software, but the authors hoped the principles could be applied to electronic systems. In order to participate in collective browsing, users have to provide details about themselves. In this study, their reluctance to give up this element of their privacy outweighed any benefits they perceived in the scheme. In fact, with refreshing honesty, Dave Nichols bravely described how this particular piece of their research had produced little data, and had thus in itself highlighted incorrect assumptions in the research paradigm. There is a long-heard complaint that researchers only report the successful, not the unsuccessful studies, thereby missing the chance of sharing the 'learning experience', so perhaps such reports are to be encouraged (though maybe within workshops rather than plenary paper sessions).
Other papers examined the effectiveness of particular electronic library systems - an important issue for library managers, who increasingly have to choose between alternative technologies for delivering the same information. Ann Morris (Loughborough University) described the FIDDO project (Focussed Investigation of Document Delivery Options) which will evaluate a number of electronic document delivery services.
Christine Dugdale (University of the West of England), in a paper based on evaluative research carried out during the ReSide Electronic Reserve  project challenged the assumption that electronic libraries necessarily provide equality of access to their users. While the electronic library overcomes many of the limitations of paper-based resources, particularly in the type of high demand material included in the ReSide system, the ability of users to access material can be affected by a number of factors. These include the IT infrastructure, the level of expertise and training of the individual user, and copyright restrictions (which may bias a collection towards resources where no copyright problems exist).
The issue of copyright was, as always, a recurring theme at the conference, and the subject of one of its workshops. In the absence of an explicit legal framework for electro-copying of copyright material, negotiations between publishers and libraries and authors still tend to be on a case-by-case basis and limited to the duration of a particular project. Looking on the hopeful side, there is a growing body of experience and examples of good practice in this area. Leah Halliday (University of Stirling) reported on the negotiations with publishers on behalf of the Scottish Collaborative On-demand Publishing Enterprise (SCOPE), which is developing a resource bank of teaching and learning materials for a consortium of 13 Scottish higher education institutions. A SCOPE  model contract has been developed. The related issue of charging was also discussed: how, and whether, the various costs should be passed on to library users. George Pitcher (Napier University) discussed the security and reporting aspects of the SCOPE system, features which can reassure publishers that their material is properly protected and managed. The system includes encryption of documents, user identification, logging of transactions, and "watermarks" on all printouts, giving details of the individual printing the document, and the date. Documents are stored in PDF format, which publishers generally seemed to prefer because of the protection it offers against copying and altering. There are additional complexities for some projects dealing with different materials, as Elizabeth Lyon's paper on the PATRON project showed: their negotiations also had to cover performing and recording rights.
Quality was another frequently mentioned term. Biz/ed  (Catherine Sladen, University of Bristol) is a subject-based information gateway (SBIG), which provides a range of users (14-19 year olds, higher education students, teachers and lecturers) with links to quality assured Web-based sources of Business information. Resources are inspected and catalogued by a distributed team of academic subject specialists from different institutions throughout the country. To be included in the gateway, resources have to meet five quality criteria: scope, content, form, process, and collection management. Biz/ed is an example of a project which has benefited from collaboration. It has close links with a number of other SBIGs, sharing the same software (the ROADS database).
The conference also covered a range of other technological ideas and developments. Michael Emly (Leeds University) discussed the technical issues raised in the Internet Library of Early Journals project  , which is producing a digitised version of a number of widely used 18th and 19th century journals. Pages are scanned, then OCRed. Scanned pages can then be retrieved by free text searching of the OCR text. The OCR text contains errors, but these are partly overcome by fuzzy search algorithms. Some improvement in accuracy could be obtained by completely re-keying the journals, but the chosen method is the most cost effective.
Two papers described strategies to combat problems caused by the huge scale of the WWW. The volume of material now accessible on the WWW means that if users want to retrieve information from one of the large WWW search engines (like Lycos) they need to be able to carry out complex queries. Gary Mooney (De Montfort University) described the application of fuzzy logic and user modelling to assist users in this. A fuzzy modelling query assistant was developed, which gathers information about the user and their needs, and then builds an appropriate query. John Kirriemuir (UKOLN, University of Bath) described the techniques of mirroring and caching which aim to reduce bottlenecks on the WWW by creating copies of heavily used resources. These raise technical issues, particularly when copying complex sites containing indexes and dynamic pages and also issues of responsibility and rights. The paper included recommendations for a service level agreement between an originating site and the mirroring site.
Christian Schloegl (Karl-Franzens-Universitat Graz, Austria) described a prototype of hypertext library catalogue (HyperKGB), which provides users enhanced browsing capabilities, and could be offered as an additional option to the traditional query-based method of searching. A number of studies (including MUMLIB, mentioned above) have suggested that most users prefer browsing as a means for retrieving information, so more intelligent browsing facilities are clearly to be welcomed.
I should also mention the other ELVIRA delegates, who ensured that the discussions which followed from the papers were lively and informed - an important ingredient to the success of any conference. With most other information available on the Web these days, it is good to know that stirring from your desk can still bring non-digitisable rewards. A wide range of organisations was represented, as well as several overseas countries, though there were not many people from public libraries - something which can hopefully be rectified next year.