This was the first workshop in MODELS Phase II, although it continued to build directly on the results of the five workshops in Phase I. A Distributed National Electronic Resource? involved the largest number of participants to date. Around 50 people gathered at the Stakis Hotel in Bath on 5-6 February 1998, to discuss management and access approaches to the growing mass of currently unconnected national resources provided by libraries, data centres, archives, subject gateways, electronic journals, clumps and others. We were grateful to Richard Heseltine of University of Hull, for once again chairing a workshop which posed a number of challenging management questions. Funding to support the workshop was provided by eLib.
The workshop did not attempt to address the thorny problem of what constitutes a Distributed National Electronic Resource (DNER). Within a higher education context, the Content Working Group of the Committee on Electronic Information (CEI) has been examining the issues and CEI itself has discussed the DNER at recent meetings. Rather than content, MODELS is instead concerned with strategic planning for achieving a managed environment, together with the technical infrastructure which will provide its basis. However the DNER envisaged has a truly national character, and is certainly not restricted to HE.
There was a distinct shared recognition at the workshop that the management approach proposed by the MODELS architecture (which has been developed via previous workshops) is the direction to take, in order to provide fully integrated end-user services, in place of the current set of stand-alone services.
A series of short presentations on the first day of the workshop set the scene, identified requirements and revealed some of the building blocks that are being put in place. The group breakout sessions identified a range of organisational and technical barriers to achieving a DNER and matched the draft MODELS Information Architecture (MIA) against existing services, to see how well it fits. The morning of the second day is usually when the pieces of the jigsaw start to come together and suggestions for ways of moving things forward emerge. A series of recommendations was produced, which is available in the full workshop report. However the fundamental issues that emerged centred around the following areas: 'information landscapes'; the move away from current attention to user interface level; market 'pull' as a force for change; and of course MIA. Given the close links between the successive MODELS workshops, it is usual for some of the same issues to re-emerge, albeit in different guises.
The concept of information landscapes bears different interpretations, but from a user point of view it essentially means that instead of being offered a mass of undifferentiated resources and services, a personal view is presented, based on the individual user's information needs. This view, or landscape, may be constructed in different ways, and initially it is likely to be fairly simple, perhaps just a set of web links. Within a MIA context, the landscape is provided by a broker and will provide much richer functionality. Landscape creation will involve three components:
- user profiles containing information about eg preferred interface, research interests
- collection/service descriptions
- mechanics of searching (eg Z39.50 Explain)
There is currently no clear solution for implementing the second component (collection/service descriptions), although the first and third are reasonably well understood, if not yet widely implemented. It is desirable for libraries to have some degree of control over user profiling.
It was acknowledged that services currently compete on user interface – people tend to distinguish between databases by the interface and become accustomed to using one particular service provider. We need to move to a situation where services also compete on server interface - in a MIA context it is also necessary for libraries to have access to the underlying search engine (via Z39.50); services that cannot provide this in future will not be able to 'plug in' to a distributed library service.
Now that the library and information environment is more mature, the issues that MODELS is raising (eg lack of collection descriptions) are becoming more real and consequently the need for solutions is more pressing. However solutions will not come about unless there is a market 'pull'. The more informed the market becomes, the more it becomes a force for change. Therefore wide dissemination of MODELS results is essential.
MIA has been developed progressively over the life of the project, with the input of a wide range of MODELS participants. It represents a high-level blueprint for how distributed information services should be designed. The library broker (or trading place) model lies at the core of the design.
The next MODELS workshop, to be held in May 1998, is specifically addressing the deployment of MIA. The overall aim of the workshop is to 'concretise' MIA - to take it one stage further into something practical, which can be used by the various parties involved: system vendors, information providers, and libraries. We need an 'embodiment' of the architecture, so will be looking to develop demonstrator scenarios.
Over the coming months, UKOLN will be taking forward the DNER workshop recommendations, together with eLib, JISC, LIC and other partners. A range of complementary dissemination activities will be undertaken, particularly to raise awareness of MIA in the relevant communities.
The full MODELS 6 workshop report, together with the programme, participants list and other supporting material, are available from: URL: http://www.ukoln.ac.uk/dlis/models/models6/