In centuries past publishers and libraries had well-defined roles and responsibilities in the dissemination of information: publishers produced publications, libraries bought them and preserved them in their stacks. In the digital world, that division of labour no longer applies, especially with regard to e-journals. Now, publishers retain ownership of the content and license access rights to libraries.

This leaves research libraries in a vulnerable position: their dependence on electronic information is growing: e-journals have come to dominate academic literature. Although publishers assure perpetual access rights to the content purchased in their licensing agreements, the question still remains who takes responsibility for preserving this electronic intellectual output.

The question is all the more pressing as digital information is fragile and very much dependent on a properly working technical environment — which must function throughout the life-cycle of the digital object. Many threats to such continuity may be identified: hardware and software will eventually become obsolete, publishers go out of business, or their access platforms may fail.

One group of libraries would seem to be specifically well-placed to assume the role of securing access to the records of science: national libraries. National libraries have a long tradition of preserving national deposit collections and their remit specifically includes the long-term perspective so needed to secure permanent access. All that would be needed is to extend this remit from printed collections to digital publications.

The Koninklijke Bibliotheek, National Library of the Netherlands, was one of the very first cultural heritage institutions to become aware of the emerging importance of digital resources. As early as 1998 it concluded an agreement with the Dutch Publishers Association to extend the Dutch voluntary deposit scheme to off-line electronic publications (cd-rom’s etc.), and in 1999 a tender was issued for the development of a long-term storage facility for electronic information resources. As no ready-made commercial products were available at the time, the KB embarked on a joint project with IBM to develop the Digital Information Archiving System (DIAS). The so-called ‘e-Depot’ became operational in January 2003 and it was the first storage facility specifically designed to store and maintain digital objects not just for a few years but indefinitely, in line with the remit of the Koninklijke Bibliotheek as the national deposit library.

Originally, the e-Depot was designed to preserve the electronic publications of the Dutch publishers, in agreement with the Dutch voluntary deposit scheme. In June 2005 an agreement was signed with the Dutch Publishers association to secure the deposit of on-line digital publications as well. Members of the association were found prepared to deposit all digital publications with a Dutch imprint. The KB in turn took on the obligation, within its deposit remit, to preserve these publications in the e-Depot. As was the case with the printed deposit collection, access to digital publications is restricted to on-site authorised users only (open-access publications being freely available online).

Some of the first archiving agreements were signed with major scientific publishers based in the Netherlands, such as Elsevier and Kluwer. As these are international publishers, the question soon arose how digital resources which are simultaneously published all over the world, fit into traditional national deposit schemes. The answer was simple: they do not.

The KB decided that a new international framework would have to be developed to preserve digital publications for the long term. As such a framework does not come to be overnight, the KB took a first step by opening up its own e-Depot facilities to all digital resources published by Dutch international publishers and indeed to all major international scientific publishers. Quite a few publishers have meanwhile concluded archiving agreements with the KB, i.e., Elsevier, Springer, Blackwell, Oxford University Press, Taylor & Francis and Sage. As of December 2008, the e-Depot has ingested more than 12 million digital objects.

Considering the costs of the infrastructure and the complexity of sustainable storage, the KB has since decided to use the e-Depot for a wider variety of services. The e-Depot now also preserves masters resulting from major Dutch digitisation programmes, the contents of the Dutch institutional repositories and the Dutch national web archive.

It soon became clear that the e-Depot was not just another book storage facility that could be added without any organisational implications. Quite the contrary: new digital workflows had to be designed and as digital preservation was such a new topic, it was immediately clear that a fair amount of research and development had to be organised as well. The KB decided to create two distinct departments in support of the e-Depot: an operational e-Depot department embedded within the Acquisitions and Processing Division, and a research and development unit within the Research & Development Division.

By embedding the operational e-Depot within the Acquisitions and Processing Division, the KB underlined from the start the parallels between digital and printed workflows: the tasks to be carried out were essentially the same (acquisitions, description and storage) — only the methods differed. Rather than dealing with individual items, as was customary with printed materials, most digital resources, especially those originating from international publishers, were processed in bulk and automatically. Obviously, bulk digital archiving workflows had to be designed for this.

As a service to smaller Dutch publishers, the KB also implemented an online web interface (the so-called ‘webloket’) which allows for individual monographs and journal issues to be deposited. The next step will be to design workflows for a growing number of diverse materials and for complex objects such as websites and other multimedia objects.

As obvious as the implementation of bulk workflows may seem from a strategic point of view, it was to be expected that workfloor practices did not adapt as easily. Cataloguers who had for years taken pride in the quality of the descriptions they produce with quite a lot of manual effort, did not readily warm to the idea of automatically generated metadata, and the fragility of the digital media and data themselves worried them as well. These staff concerns were of course quite justified, but with the influx of millions of digital publications it became quite clear very soon that manual cataloguing was just not a feasible option. Six years hence it is encouraging to see that workflows for printed and digital materials are converging and that staff are in fact benefiting from each other’s expertise. The research team, in turn, greatly profits from the experiences of the day-to-day running of a digital archive.

The Operational e-Depot Department is staffed by ten full-time employees, six of which are collection managers. They are responsible for processing incoming data. Elsewhere in the organisation, three employees provide IT support. The Digital Preservation R&D Department comprises six full-time staff. These numbers do not include temporary staff for international projects and the present programme to set up the next-generation e-Depot (see below). Other facts and figures are available from [Addressing the Future of Preserving the Past, 2007, p. 45].

The KB’s e-Depot has now been in operation for six years, and the invaluable experience gained in these years, both in the day-to-day running of the e-Depot and in conducting research and development, now enable us to evaluate our position and the impact of the e-Depot on our entire organisation.

The National Library of the Netherlands firmly stands by its decision to operate this e-Depot and to open it up to international publishers. Such degree of commitment is vital for an initiative which is still by no means routine. We can also ascertain that the e-Depot has become a driving force for renewal and change within the organisation. This influence is particularly visible in three areas:

First of all, our collections have been influenced in a major way. We now preserve almost two digital objects for every printed object in our collection. What is more, the provenance of these digital objects is distinctly international. Indeed, the world’s records of science find their way to our collections.

Secondly, the e-Depot has had a major impact on the technological infrastructure of the library, including changes in metadata modelling and handling. More traditional library services are profiting from these changes as well.

Thirdly, running the e-Depot has affected our people and the organisation. It was not an obvious management choice to set up two distinct departments, one for the operational e-Depot and one for research and development. This division of labour required high levels of coordination between departments as well as sound quality assurance and knowledge management. However, after six years, we can say it has been worth it. It has enabled the digital preservation research team to focus upon research issues and take an active role in many international projects (PLANETS, DRIVER, KEEP), while not losing touch with a firm practical basis and solutions which can indeed be implemented. Embedding the operational e-Depot Department within the Acquisitions and Processing Division has brought the best of two worlds together: frequent interchanges between the rich experience of two centuries of library management and a new workforce with an international, state-of-the art orientation and new digital skills, benefit the entire organisation.

On a technical level, the core of the e-Depot is the Digital Information Archiving System (DIAS), which was developed between 2000 and 2002. DIAS is designed according to the OAIS reference model to perform functions such as ingest, archival storage, data management and dissemination [Reference model for an open-archival information system, 2002]. Developing the e-Depot was a truly pioneering effort at the time, as there were no existing models or tools for quality assurance.

During the first six years of the e-Depot’s operation, the focus was on processing e-journal articles. Generally these are objects of a similar type and format. The workflows designed to accommodate these e-journals clearly worked, as over 12 million e-journal articles were ingested since 2003. However, the e-Depot is anticipating a future influx of many more different types of content from a variety of content suppliers, and the present technical environment is clearly not suited for such additional tasks. The architecture of the system requires major redesign, the infrastructure needs to be scaled up to be able to process larger amounts of objects of a wider variety and newly developed tools for quality control need to be implemented. Thus, early in 2008 KB embarked upon a large-scale programme directed towards the overall upgrade of the e-Depot.

As time went on we realised that improvements on the existing system are extremely difficult to put in place. The e-Depot is clearly a system of its — pioneering — time, and since then much has changed in the way we think about how such systems should be designed. The KB has therefore decided to start working on a completely new e-Depot system, the next generation e-Depot. Collaboration with the other users of the DIAS system, the German National Library and the State and University Library of Göttingen in this effort was a matter of course. But quite recently the KB has also reached out to other national libraries.

The challenges posed by permanent access are quite daunting, both in a technical and in an organisational/financial sense. No single institution or even country is capable of solving the many inherent problems single handedly. Moreover, responsibilities are diffuse [Long-term Preservation, 2008]. In order to deal with these enormous challenges, national and international cooperation is called for. From the start, the KB actively sought cooperation. The digital preservation research department participates in a number of major European projects designed to facilitate permanent access to digital objects, such as PLANETS, DRIVER and, lately, PARSE.insight and KEEP (Keeping Emulation Environments Portable).

But these projects focus mainly on technical issues: development of tools, services and preservation strategies. In addition, they are mostly of a temporary nature: once the projects are completed and funding ends, there is a risk that the results disappear into cyberspace.

But digital preservation is a long-term game, and definitely not just a set of technological problems. It involves the grander problem of organising ourselves over time and as a society. Digital preservation is also about selecting what materials should be preserved, and in what form (social and cultural issues); what rights are needed to support permanent access (economic and legal issues); who is preserving what (responsibilities). Digital preservation is an ongoing, long-term commitment, shared and met by many stakeholders.

There is a great need for long-term, sustainable partnerships with corresponding funding to take up where temporary (technological) projects leave off. The KB has taken the lead in developing such partnerships at several levels:

An example of cooperation on a practical level within The Netherlands is the network of Digital Academic Repositories, in which the National Library fulfils the role of safe place for academic output stored in institutional repositories. The participants retain responsibility for and control over their own data and provide access to them, while the KB takes responsibility for storage and long-term preservation, thereby enabling the universities to concentrate on their research work. This model is used in a wider perspective within the European DRIVER project.At the national level the KB participates in the cross-domain Netherlands Coalition for Digital Preservation (NCDD) which rallies major stakeholders from the research community, government, and cultural heritage institutions around a joint strategic agenda for permanent access to digital information in the Netherlands.

In order to coordinate various national initiatives, the KB and the British Library initiated the Alliance for Permanent Access, in which major research organisations such as ESF and CERN join forces with the International Association of STM Publishers and national coalitions to develop a sustainable infrastructure for research data. Partners include organisations from the Netherlands, the United Kingdom, Germany, Switzerland and the United States.

On a world-wide level the KB promotes the idea of a Safe Places Network of trusted digital repositories which share the responsibility for keeping digital publications safe on a global scale. The KB is of the opinion that continuous research and development efforts require substantial financial, technical and staffing commitments that exceed the possibilities of individual institutions. As a consequence we expect that only a limited number of institutions will commit themselves to permanent archiving of the records of science. A network of such ‘Safe Places’ would ensure a more systematic and concerted approach to digital preservation of scientific information. The network would also fill the gap left by national deposit arrangements. It is envisaged to be a network of a limited number of institutions with trusted digital archives which will collaborate to ensure that the ‘records of science’ published by international publishers are permanently archived and continue to be available to future generations. The collaborative effort will be geared to sharing the responsibility for complete, world-wide coverage and allocating tasks between participating institutions.

As this sharing of costs and technology, best practices and common standards must include all stakeholders in the digital life-cycle, the KB favours the inclusion of publishers in such a network.

The KB e-Depot has now been in operation for six years. During that time we have acquired lots of experience and knowledge. In addition, years of research and development have started to pay off. For instance, the KB Research and Development Department produced a clear list of functionalities which are being implemented in order to ensure ongoing durability of the ingested materials.

But the rate of change in ICT, in publishing, in libraries and in preservation is rapid and the scale is enormous. Thus this is not the time for us to look back and be satisfied. Rather, new challenges come on our path every day and we are willing to take them on. Ensuring permanent access to digital objects is a complicated task which requires collaboration on a national, a European and a worldwide scale. It is hoped that more organisations and countries will join the existing initiatives to create a sustainable infrastructure to support permanent access to digital information on a global scale.

Alliance for Permanent Access, http:/www.alliancepermanentaccess.eu

DARE, Network of Dutch Digital Academic Repositories, http://www.narcis.info

DRIVER, Digital Repository Infrastructure Vision for European Research, http://www.driver-support.eu/en/

e-Depot of the National Library of the Netherlands, http://www.kb.nl/dnp/e-depot/e-depot-en.html

KEEP, Keeping Emulation Environments Portable, http://cordis.europa.eu/fetch?CALLER=FP7_PROJ_EN&ACTION=D&DOC=1&CAT=PROJ&QUERY=011f37a73b31:61ba:091d22f8&RCN=89496

NCDD, Netherlands Coalition for Digital Preservation, http://www.ncdd.nl/en/

PARSE. Insight, Permanent Access to the Records of Science in Europe, http://www.parse-insight.eu/

PLANETS, http://www.planets-project.eu/