The JISC-funded PORTAL Project  examined and established which services users wished to have made available through an institutional portal. The results of this project have provided firm guidance to institutional portal developers in planning the services they wished to present. In particular, there was common demand amongst users for access to library-based services and resources within a portal environment. Portal technology developments at the time of the PORTAL Project were not, unfortunately, at a stage that allowed full testing of the findings from this research. Technologies have now matured to allow the investigation of how users wish to see the services they requested presented to them in portal and non-portal environments and in which contexts. The Contextual Resource Evaluation Environment (CREE) Project  has been funded by JISC to carry out this investigation, focussing on the library-based services that generated so much interest.
The CREE Project is carrying out a mixture of user requirements evaluations and technical investigations, including user testing against services presented in different contexts through demonstrators. JISC has previously funded the development of a number of portal services and other tools to support access to library-based resources ; CREE is providing a platform to test a range of these with users in different environments. The technical development strand will make use of standards that have emerged since the PORTAL Project (JSR 168  and WSRP ) to enable the delivery of such services through a portal framework. There has been much discussion on how specific library-based search tools might be used alongside more generic Internet search tools, and CREE will also test this relationship through the inclusion of Google in the demonstrators.
The emergence of the JISC Information Environment  has occurred in parallel to significant interest in, and adoption of, portal technology across the Higher Education sector and beyond. Each development addresses, in part, similar questions: how do users relate to, organise and work with the vastly increased number of resources and information systems available to them? How does an organisation ensure that the substantial investment they make in these services achieves the maximum possible take up and use? Portals - of whatever type - seek to address these questions by presenting the user with a coherent view of disparate resources, data and applications.
Much within this emerging landscape remains unknown. The user is frequently faced with a contextual 'jump' which is driven by limitations in the systems themselves rather than by user need. These 'jumps' may range from institutional portal to VLE (Virtual Learning Environment) to library system, and to the broader range of subject, media and community portals and other search tools available in the Information Environment and externally. For library-based services it is uncertain where it will be most valuable to present relevant search tools: should they be delivered through the institutional portal, a VLE, the library Web site, or a combination of these? Little evidence is available to support how users might want them to appear in practice in the daily contexts of learning, teaching and research.
Standards have recently emerged which may provide technical solutions for these 'jumps'. A Java Community Process standard, Java Specification Request (JSR) 168, describes a common method of rendering a 'portlet' (a portal component sometimes also referred to as a 'channel') within a Java-based portal framework. JSR 168 complements the Organization for the Advancement of Structured Information Standards (OASIS) Web Services for Remote Portlets (WSRP) standard, which enables remote application functionality to be accessed and used from a 'local' portal. The portal vendor community is rapidly adopting JSR 168 and WSRP and this is providing an impetus to develop services making use of these standards.
Taken together, these standards potentially offer a great deal of flexibility in the integration of nationally and locally provided services. They suggest a technical means of providing integration to portals within the Information Environment and to an institutional portal in particular. The work undertaken as part of the PORTAL Project generated considerable data regarding which services users wished to access. CREE seeks to use these emerging standards and build on this evidence by extending it to the domain of how users wish to access and use library-based services, and where those services might best be presented in a series of specific and generic demonstrator contexts.
CREE aims to:
It is a key objective of CREE that the results of these activities are disseminated effectively to both the HE/FE community and relevant standards bodies. The project Web site  will act as the conduit for this dissemination and will be updated regularly.
A consortium of partners from across Higher and Further Education has been created to take these aims forward through two parallel strands of activity covering user requirements evaluation and technical development. These will each produce results in their own accord, but are also linked to enable user evaluation against systems that users can see and use for real.
The University of Hull has been delivering services through its institutional portal since September 2003. The portal is built around the uPortal framework  and was established as part of the University's Digital University initiative  to develop e-services to support the business processes of the University. Hull's work in this area has been rewarded with Sun Centre of Excellence status . This experience with uPortal will be used to develop the demonstrators for user testing. It is the intention of the JSR 168 and WSRP standards that functionality can be presented in any conformant portal, and Hull will carry out interoperability testing within the Sun JES Portal  and tested against Oracle's Portal Verification Service  to examine this.
The user requirements evaluation work within CREE is also being co-ordinated through Hull, working with partners at Newark and Sherwood College and the University of Oxford, representing Further and Higher Education, to gather as wide a range of views as possible from different constituencies.
The library-based resources being adapted for use within the demonstrator(s) have been developed through JISC and other publicly-funded projects in recent years. All have proved themselves to be robust and valuable services that are now being used in a variety of environments in their own right. They are thus ideal to enable a comparison to be carried out within portal and non-portal environments, as well as being prime candidates for testing the practicality of the JSR 168 and WSRP standards.
In addition to examining existing tools, CREE is also seeking to test out how these tools might be presented alongside more generic Internet search engines. To enable this, Google is being adapted for use within a portal environment by instructional media + magic (im+m) , a US consultancy and software development company that has been at the forefront in the development of uPortal and contributed much to this. The use of XSLT (Extensible Stylesheet Language Transformations) to assist with presentation of resources through a portlet will also be examined by im+m.
When users interact with the range of systems and services they encounter within an institution, they are guided to these through a variety of routes (see Figure 1). The factors that determine these routes are both organisational and technical. Institutions commonly have a library catalogue and quite often a VLE; many are examining or are running a portal at either the institutional and/or library level. Direct access to external and internal services is commonplace through a myriad of interfaces, and internal systems such as the VLE often need to interact with these. Existing methods to deliver these systems and services are often proprietary and limit the choices organisations have to build services around them; processes are too often built around what the system can do rather than the other way round. The use of open standards can provide flexibility and choice in the way organisations, and individuals within them, access services. However, such technical solutions are only valuable if they are guided by evidence of how users would like to see services and information presented to them - getting the systems to do what the user needs. This evidence can be both theoretical, gathered through asking users what they want, and practical, through showing users what possibilities there are and gathering their feedback.
CREE is addressing both the theoretical and practical angles. A national user survey  is examining how users currently search using existing services and gathering user opinion on suggested methods for integrating these within local systems. Of particular interest here are links into the VLE, in order to support the learning process more fully and enable integration of library and administrative systems within this. The findings from this survey will be tested and expanded through a series of focus groups with both staff and students within both Higher and Further Education institutions. Combining evidence from these two sources will then underpin the development of demonstrators to allow users to see what might be possible and provide feedback on this. The results of this evaluation will provide evidence to assist with the effective use of technologies in delivering services to users through both portal and non-portal environments.
The evaluation will be given further depth by analysing user needs and behaviours across one disciplinary community. The historic environment sector is a microcosm of the problems faced by portal developers. There is a broad range of organisations that require access to information about the historic environment, and in turn produce information about it, from national agencies to local public archives and museums to universities and schools. Across these there is huge and under-exploited public interest. So, like many datasets used in teaching and research, services often come from outside Higher Education, and are of much wider interest than to universities and colleges alone. CREE, in partnership with the Historic Environment Information Resources Network, is aiming to take a snapshot of user needs for research data services right across this spectrum of interest. The national user survey of Higher and Further Education users will be adapted for a national survey of those using historic environment information services (e.g., Archsearch  and the British and Irish Archaeological Bibliography ). The differences and similarities between the results across the two surveys will make interesting reading.
Technical development of the demonstrators to assist with the user evaluation has provided an opportunity to carry out a detailed examination of the standards produced recently to facilitate the delivery of services through a portal. JSR 168  and WSRP  largely aim to achieve the same end, the presentation of services through portlets within an overall portal framework. JSR 168, being a Java standard, specifies a route for doing this with Java-based portals, and many of the portal frameworks currently available, including uPortal, Sun JES, OracleAS , IBM WebSphere  and BEA WebLogic  fall into this category. WSRP, following a Web services path, is platform agnostic and can be used to present services through any WSRP-conformant portal, and many of those listed above also support this in varying degrees.
The two standards could be seen as competing with each other on this basis, both offering similar routes to presenting services but differing on the specificity of the portal framework platform. There is, though, also a difference of focus in where the service or tool being presented is located, locally for JSR 168 or remotely for WSRP. To assess fully the possibilities of using these standards to build the demonstrators it was thus decided that portlets using both standards would be developed for each tool. Initial investigations  by the CREE partners examined the requirements for using each standard and how this related to their existing tool(s): JAFER and HEIRPORT are Java-based already, whilst Xgrain and Balsa are written in Perl. This initial investigation revealed independently at each partner that from a development viewpoint JSR 168 has proven easier to work with at this stage. Factors such as the software being used to build the portlets, the requirements of the standards, and the requirements of the main testbed portal framework being used within CREE, uPortal, have contributed to this finding. For the Java-based tools, this was a logical choice in any case. Interestingly, it also proved to be a logical choice for the Perl-based tools as there is no current toolkit to enable Perl tools to be presented as WSRP portlets.
There is, however, an existing tool to enable Java tools to be presented as WSRP portlets, WSRP4J . This toolkit can be used with both Java tools and JSR 168 portlets. Developing a JSR 168 portlet can thus be used as a standard way of presenting Java-based services to facilitate the development of WSRP portlets using the WSRP4J toolkit (see Figure 2). Using this route, both standards can be used and maximum exposure can be gained through the full range of conformant portals. As the services being examined within CREE will be accessed remotely for the most part, this path to WSRP potentially offers great benefits for the project.
A specific case of how JSR 168 and WSRP can be used in this way is demonstrated by the development of JAFER. Each JAFER service (Z39.50, SRW, Distributed search) is being adapted as a JSR 168 portlet. WSRP4J is then being used to present, or 'wrap' these as WSRP portlets (see Figure 3).
It should be noted that WSRP and emerging JSR 168 best practice, tested through work carried out within CREE by im+m, enables control of portlet look-and-feel within the portal framework by style sheet transformations applied by the host portal. This has implications not only for 'branding' issues, enabling visual integration and consistency, but also for accessibility for individuals with disabilities, for whom specific presentations of the portlets can be designed.
The CREE Project, through its user requirements evaluation and technical development work, aims to provide a body of evidence that can support the development of services for presentation through a portal. Beyond this, CREE is more widely seeking to discover how users wish to interact with services in the context of their learning, teaching and research. There is likely to be no one answer to this, as individuals will have individual requirements. However, identifying trends and preferences will enable the technology and standards to be used in a flexible manner to meet user, and organisational, needs.