ALIA home | Online home | search | papers | keynotes | program | sponsors | satellite events
|
Edward H T Lim, Assistant Project Director, AARLIN, La Trobe University Library, e.lim@latrobe.edu.au
Earle Gow, University Librarian and Project Director, AARLIN, La Trobe University Library, e.gow@latrobe.edu.au
The AARLIN project, supported by the Australian Research Council, is a collaborative project involving a number of Australian universities and the National Library of Australia. Its principal aim is to develop a structure and framework for providing unmediated, personalized and seamless end user access to the analogue and digital resources of Australian university and research libraries from workstations of research staff and students.
This paper outlines the outcome of the Pilot project, which involved six of the participating universities and their academic staff and researchers, and also deals with the implementation phase, which has received $2.8 million funding under the Federal Government's Systemic Infrastructure Initiative. The technologies underpinning the project are discussed briefly. They include standards for interoperability, the use of LDAP directories for authentication, the protocols needed to link the portal to local authentication systems, the use of 'pull' and 'push' technologies, cross domain searching using a variety of search protocols - Z39.50, http, SQL, and OpenURL linking.
Finally, the paper deals with the advantages and disadvantages of such a collaborative approach as well as the legal and administrative framework that will ensure the long-term viability of this collaboration.
The proposal to develop an Australian Academic and Research Library Network (AARLIN) emerged from a specially convened meeting of the Council of Australian University Librarians (CAUL), held in Hobart in March 1999. The aim of the meeting was to develop a strategic plan for CAUL, which would be implemented over the period of 1999 - 2004. At the time, academic libraries were experiencing difficulties in meeting the needs and demands of the scholarly community, particularly in regard to the information infrastructure to support research and teaching informed by research. These concerns were shared by all Australian universities and were raised in a number of public forums on the 'crisis' in scholarly communication. CAUL had been pursuing collaborative resource sharing initiatives and had supported projects aimed at improving access to scholarly literature. However, progress was slow and ill-fated projects, like JANUS, were unhelpful. Independent reviews of the national information infrastructure confirmed the parlous state of disrepair, the consequences of which had serious long-term difficulties for the national research base.
At the same time, scholarly publishing, (and the term is used generously) was undergoing significant change. The advent of the Web and related advances in information and communications technology afforded opportunities for a radical rethink of approaches to the publishing and distribution information, including scholarly writing. Some saw this outpouring of information of variable quality as having a tsunami effect. A solution had to be found. The problem had become global.
In the United States the harvesting of Web publishing became an issue for those involved in developing systems and services aimed at discovering and providing prompt access to scholarly information.
As Sarah E Thomas (2000) phrased it, 'the world's information resources are abundant, but time is a scarce commodity. The ideal discovery tool, therefore, is one which consults omnivorously, but which returns a selection of relevant results in quick succession .... Such a tool is still imaginary, although it could become a reality in the near future if librarians organise themselves appropriately and commit the resources to design it'.
In September 1999 a strategic issues forum for directors of academic libraries in the United States, hosted by the Association of Research Libraries and OCLC, advocated the creation of a 'scholars portal' (Campbell 2000).
It was reassuring to those involved in developing the AARLIN proposal that their work had global implications. In broad terms the CAUL group set out to research and develop an architecture using portal technologies, which would integrate the various library resource sharing services so that users would have an integrated view of the information resources available in collaborating libraries. Importantly, the portal would be able to provide access to information in a range of formats.
Following a joint meeting of the Executive committees of CAUL and the Council of Australian University Directors of Information Technology (CAUDIT), the development of AARLIN became a joint CAUL/CAUDIT initiative in late 1999.
It is one thing to generate exciting and challenging ideas, another to gain the support and resources required to translate these into an operational system.
A campaign of rallying and seeking support for the project was pursued, culminating in the successful application for funding from the Australian Research Council's RIEF (Research Infrastructure, Equipment and Facilities) Scheme in 2001, supplemented by contributions from 19 university libraries and the National Library of Australia. This was more of a 'proof of concept' project to develop and test the functionality and viability of the AARLIN facility. Expressions of interest were sought from vendors of appropriate software to enable the AARLIN pilot project to proceed. Ex Libris was selected and six universities agreed to become test sites to trial the AARLIN prototype.
Academic staff and research students from each university test site were trained to use AARLIN. Valuable feedback, through pre and post AARLIN trial questionnaires, was most encouraging. Equally encouraging was the subsequent receipt of a $2,823,900 grant from the DEST (Department of Education, Science and Training) Systemic Infrastructure Initiative Scheme, to continue the development of the prototype into an operational system that could be rolled-out to those university libraries that wished to participate.
It is anticipated that AARLIN Phase 2 will commence early in 2003 with 22 university libraries, covering all States and Territories of Australia, confirming their participation.
The AARLIN service model is based on a consortium model for resource sharing and utilises software that supports multiple institutions rather than a single institution. The key to the success of the project is seen to be collaboration. In this context, collaboration should be seen to be quite different from co-operation. Brian Hawkins (2000), quoting Andrea Youngdahl, states that there is a major difference between co-operation and collaboration. co-operation involves some degree of informal sharing but allows participants to continue to operate fairly autonomously. Collaboration, on the other hand, 'involves a synergistic - not an additive - solution.' It includes 'commitment and investment of resources, based on a shared vision.' Hawkins further claims that for institutions of higher education to be viable in the future, collaboration is absolutely necessary. The new electronic environment requires that the higher education sector, including libraries, should rely on new forms of collaboration not only among themselves, but also with external partners, to share information resources electronically, to break down current organisational structures, and to reject the 'go-it-alone' mode that higher education libraries have for so long pursued. In a sense, therefore the AARLIN collaborative model is a unique implementation for Australia. For the first time, it embraces a model where institutions no longer have complete autonomy to do what they like in terms of information delivery to researchers, and where their services are intricately intertwined.
AARLIN uses portal technology to provide a whole host of services for researchers, including:
As shown in Fig. 1, the AARLIN Service model is built around a national portal framework, which is linked to the local authentication systems of the participating universities. When a user logs on, the portal ensures that the user is an authorised user by communicating with the authentication system of the user's institution. The authentication service passes to the portal metadata relating to the 'user's profile' - which may include information such as the research or subject interests of the user, his/her status (e.g. undergraduate, postgraduate, academic staff), and other relevant information. On the basis of that profile, the portal will 'push' to the user a suite of relevant information resources (or information landscape). Thus different users would each view a different information landscape. Individual users can further refine their access to the relevant resources by adding or deleting individual items from their default list of resources. The provision of this 'push' facility reduces the amount of effort required by users to access relevant information, since many of them experience increasing difficulty in determining which of the many databases and other electronic information resources have relevance for them.
In linking the portal to local authentication systems, two basic protocols will be employed. The first is SOAP (Simple Object Access Protocol), which is an XML-based protocol for exchanging information in a decentralised or distributed environment. (World Wide Web Consortium [W3C], 2000). The second is LDAP (Lightweight Directory Access Protocol), which is a protocol for accessing online directory services (usually X.500-based). A recent [unpublished] survey of authentication systems by Vic Elliot, University Librarian of the University of Tasmania, on behalf of CAUL reveals that most universities build their authentication systems around LDAP. It is also clear from Vic Elliot's survey, that some universities have not yet implemented institution wide authentication systems. Where such a situation exists, the AARLIN portal will communicate with whatever authentication systems used by libraries, even those based on the patron data of local library automation systems.
Another major component of the portal is the common user interface, which allows parallel searching of a diverse range of databases, information resources and websites using multiple protocols. These protocols include Z39.50, HTTP, and SQL. Thus, it would be possible for a user with a single search query to search across multiple citation and full text databases, online library catalogues, Internet search engines, websites and subject gateways, and get a uniform search outcome from this parallel search. The advantage of such a facility should not be underestimated. In fact, a study by Cornell University's Albert R. Mann Library showed that users appreciate the availability of a common user interface to 'conceal the complexities of an information landscape characterised by numerous, disparate information resources' and the elimination of the need to master different search interfaces for different databases. (Payette and Rieger, 1997). This finding is consistent with the findings of a Monash University survey of databases conducted in 2001 in which many of those surveyed complained that access was too complicated and that multiple search syntaxes caused considerable confusion. (Huggard, et al., 2002)
In addition to being able to conduct parallel searching, the portal can pass relevant reference metadata to openlinking software using the OpenURL framework. The openlinking software will use the metadata to 'resolve' what types of extended services the user is entitled to according to the user context. For instance, it might display links to full text resources, which the user's library system has subscribed to, but not to other relevant resources, which the user is not entitled to use. Similarly, if there is no full text access, the openlinking software might display a document request form populated with bibliographic metadata, if the user is an academic staff, but will not display this form if the user is an undergraduate and is not entitled to document delivery services.
Finally, the portal will have an SDI or Alerting service, which can keep users informed of new information resources in their fields of interest.
AARLIN can incorporate a number of additional features, although these are not within the scope of the current project. These include:
As mentioned earlier, AARLIN is a collaborative project, and there are many benefits in approaching the development of a national portal collaboratively rather than individually or independently. Some of the benefits include the following:
Political. Australian university libraries have frequently (and unfairly) been accused of not being willing to co-opte or share resources. This kind of criticism, which has come from Government ministers as well as bureaucrats is not surprising in view of the competitive environment in which universities are required to operate. It is patently unfair because Australian university libraries, through CAUL and other regional co-operatives, in fact participate in many co-operative projects, including interlibrary lending, co-operative storage, reciprocal borrowing, and consortium purchases of electronic and analogue resources. Unfortunately, however, many of these co-operative projects are unknown outside the sector because of their lack of visibility. AARLIN will provide a tangible framework on which these collaboration and co-operative projects can hang. Some recent examples include national site licences, reciprocal borrowing, document delivery, co-operative acquisitions, and online reference and help services. More widespread dissemination of information about these will send a powerful political message to the Australian government in Canberra.
Financial. The financial imperatives for establishing a national system are overwhelming. It costs anything between $300 000 and $1.0 million to set up a portal individually. These include hardware and software costs only and do not include staffing costs. On the other hand, the cost of a national system works out at less than one third per institution. The costs savings are not only in hardware and software but include potentially large time and salary savings. These are achieved through collaborative work in identifying databases and websites, and in configuring targets and writing software scripts, adapters or drivers. The work involved in identifying and configuring targets is often under-estimated. By collaborating participants only need to configure targets or write 'scripts' once, and share the results among themselves.
Staffing. A collaborative project creates a positive environment and generates considerable goodwill among the collaborating staff. The AARLIN pilot has generated a lot of positive vibes among participants resulting from the sharing of expertise. There are also staff development opportunities arising from the ability to learn from one another, and from the development of individual skills that previously did not exist. A consortium approach also strengthens the participants' ability to communicate with database and software vendors about deficiencies in their implementation of standards and protocols, and to pressure them to make changes.
Scalability. One of the criticisms relating to a collaborative project of this scale is that scalability might become a problem - since the system has to provide services for hundreds of thousands of users. However, the hardware and software architecture have been designed to cope with this problem, and the following typologies are permitted:
The centralised model requires the acquisition of very expensive Sun servers, and will probably not be pursued on the ground of costs, even though a centralised model has many attractions including the fact that it is easier to retain quality control of the configured targets, to maintain the integrity of the shared data through real-time updates, and to more easily co-ordinate activities across the system.
A distributed model allows AARLIN management to locate servers in the various states, thus solving some of the problems relating to the cost of Internet traffic. At the same time, it permits AARLIN to make use of cheaper hardware, to provide a failsafe solution through the use of mirror servers, and to incrementally increase the number of servers as the need arises without suffering a cost blowout.
The third configuration combines the advantages and disadvantages of the above two options.
The AARLIN team is currently examining all the options and appears to be leaning towards the second or third options.
While there are many advantages in pursuing collaboration, there are also some disadvantages, which need to be mentioned. The first relates to the loss of some degree of local autonomy, since decisions, such as those relating to the rollout, software enhancements, or targets to be given priority, will be based on consensus. There will be also be a need to standardize workflows. A second disadvantage relates to possible delays in dealing with systems problems. Furthermore, in the current competitive environment, each university is interested in differentiating itself from its competitors by emphasing their advantages and unique characteristics. This will be lost when the universities participate in the same system.
Finally, although this may not necessarily be viewed as a disadvantage, participants will have to adopt common standards to ensure that their systems will interoperate with those of their collaborative partners.
From a technical perspective, AARLIN is concerned with how interoperability can be achieved:
For the various components of the portal to work, various standards and protocols have to be adopted. The primary ones are:
Many vendors do not implement these protocols fully or correctly, while some of the protocols (e.g. OpenURL) have yet to be fully implemented. Nevertheless, the adoption of these protocols and standards are necessary to ensure continued interoperability for AARLIN so that the transition from the current system to future enhancements will occur smoothly.
Funding for AARLIN Phase 2 will cease at the end of 2004, after which new sources of funding will have to be found. To ensure the sustainability of the system, the AARLIN planners will take steps to develop a business plan in 2003 or early 2004. A Consultant will probably be appointed to assist in this development. During the course of the project, careful statistics and records of costs will be kept so that the necessary supporting data can be made available to the Consultant.
Apart from developing a business case, the Consultant will also need to consider various organisational and administrative models to run and management the AARLIN system. Preliminary thinking indicates that three alternatives will need to be considered, viz:
In any case, the legal and administrative structures can be quite complex, and will need to be examined thoroughly by the participating universities.
The significance of AARLIN is that it takes the commitment of libraries to the provision of quality services to another level of sophistication. Furthermore, the aggregation of effort in the consortial effort provides palpable evidence that this is a way forward for academic libraries.
Libraries have been evolving. In the days when books were in closed stacks, users relied on library staff to fetch the books for them. The library operated like a retail shop, where all the articles of value were kept under lock and key or behind the counter, and delivered directly to the customer by the retailer. Libraries then moved into the era of open access or open stacks, where users could browse in the stacks, select the book or books that they wanted to read, and then check them out at the loans counter. This is the supermarket or self-service model, and it is this model that libraries continue to use today. However, this is a standalone model and does not promote collaboration easily. To remain relevant libraries have to develop a new collaborative model for providing services, preferably using e-commerce and portal technologies. The products that libraries deal with - information products like electronic documents and digital objects - have been called 'the purest form of Electronic Commerce' (Schutzer, 1996) and consequently, e-commerce technologies are particularly suited to encourage collaboration in the virtual environment.
Campbell, Jerry D. August 2000. 'The case for creating a scholar portal on the Web : a white paper'. ARL Newsletter, Issue 211,
Hawkins, Brian. 2000. 'Libraries, Knowledge Management, and Higher Education in an Electronic Environment'. Keynote address, presented at ALIA 2000: capitalising on knowledge, the information profession in the 21st century, 24-26 October 2000. [online] http://www.alia.org.au/conferences/alia2000/proceedings/ [Accessed 12 August 2002]
Huggard, Simon, et al. 2002. 'Monash Library Database Usage Survey'. In: VALA 11th Biennial Conference and Exhibition:e-volving Information Futures, 6-8 Feb 2002. Melbourne: VALA, p.313-342.
Payette, Sandra D. and Rieger, Oya Y. April 1997. 'Z39.50: the User's Perspective'. [online] D-Lib magazine. http://www.dlib.org/april97/cornell/04payette.html. [Accessed 14 October 2002]
Schutzer, Daniel. April 1996 'A need for a common infrastructure: digital libraries and electronic commerce' [online] D-Lib magazine. http://www.dlib.org/dlib/april96/04schutzer.html [Accessed 12 August 2002]
Thomas, Sarah. October 2000 .'Abundance, Attention and Access : of portals and catalogs'. ARL Newsletter, issue 212,
World Wide Web Consortium (W3C). May 2000. 'Simple Object Access Protocol (SOAP) 1.1' [online] http://www.w3.org/TR/SOAP/ [Accessed 11 October 2002]