40 minute presentation = 20 slides Cloud metaphor – getting a little ragged? Used here to suggest that the emergence of a shared repository infrastructure is, if somewhat slowly, beginning to affect the ‘climate’ in which libraries operate. Emergence of cooperative storage facilities has enabled a significant redistribution of the aggregate print collection Concentrating library resources in large-scale regional and national hubs where operational costs can be more effectively controlled Allowing individual libraries to develop service portfolios that are more closely aligned with current scholarly practice and the needs of local users. The task before us today (and tomorrow) is to examine the evolving repository library paradigm and establish some shared expectations for its future development . Since I have the privilege of speaking at the beginning of this conference, I’d like to advance an intentionally provocative claim that I hope will help to galvanize debate and discussion. . . . Advance a generic framework for understanding why, and when, individual libraries embrace cooperative management strategies. This will help to situate the repository library paradigm in a larger organizational context and suggest some ways in which we can strengthen and improve existing infrastructure. Not about space
“ Entering Finnish Airspace” photograph by wili hybrid: http://www.flickr.com/photos/wili/267263693
It is especially fitting that this conference on the ‘Universal Repository Library,’ which draws a self-conscious parallel between reference-able namespaces on the Web and a ‘global’ library service architecture, is being held in a nation (republic) that [click] recently declared broadband Internet access to be a legal right. This image of a user deep in the Finnish forest, connecting to the network on his laptop, exemplifies the expectations of ubiquitous access that are transforming the library environment by demonstrating (as Lorcan has often said) that ‘discovery happens elsewhere.’ The Kuopio-3 conference organizers have quite rightly, I think, highlighted ‘sustainability of the digital copy’ as a critical issue for repository libraries, considered both individually and collectively. Confidence in the persistence and sustainability of large scale digital collections may be the single most important factor in enabling a transformation of the traditional, locally-organized paradigm of the library. The continued growth and sustainability of large-scale print repositories will depend to a large degree on the value that contributors derive from transferring print holdings (and operations) to regional and national service centers and, secondarily, the value that libraries receive through participation in cooperative service agreements for preservation and access. I will return to this point later in the presentation, when I discuss some work that OCLC Research has undertaken to examine the interdependencies between large-scale print and digital repositories. But first, I want to make some general observations on library organization and the changing locus of collection management as a core library operation.
Contrasting views on collection management, as embodied by two iconic Finnish characters. [ click ] The Hemulen – collecting as an intrinsically rewarding personal enterprise [ click ] Little My – local ownership as a unnecessary burden Dichotomy is over-stated and cartoonish, but it has its parallel in the library environment. Not just a matter of libraries acknowledging that interlibrary loan has a permanent place in library management. Instead, a way of characterizing different organizational models of the library. On one end of the continuum, we have a model of the library that is organized around a locally managed physical inventory. The nature of that inventory may change, just as the Hemulen’s collecting interests change, but the operational model remains the same. Acquisition, cataloging, management are all managed ‘internally.’ Farther along this continuum, there is a model of the library in which management of print collections is no longer a core activity. Resources may still flow toward the acquisition, discovery, delivery and preservation of print, but these operations are no longer organized within the library. Instead, the institution may rely on cooperative service agreements with external providers (which may be other libraries) that manage collections on its behalf. In the current library environment, these two organizational models represent equally valid options – just as in the ecology of Moomin Valley there is a place for both the Hemulen and Little My. I want now to consider the factors that determine where individual libraries are situated on this continuum.
I want to say a few more words about the organization of library collections and services because I think it will help us situate the ‘repository library model’ in a larger social and economic context, and more importantly, characterize the conditions under which we are likely to see a global network of repository libraries emerge and flourish.
Ronald Coase, Nobel Laureate in 1991, published his seminal Theory of Firm in 1937. ‘ Theory of the Firm’ holds that the changing locus of economic activities (which can be either managed internally or transacted for in the marketplace) is determined in large part by transaction costs. An opportunistic model of how institutions are organized. Oversimplified here, but still useful as a framework for understanding large-scale changes in the library organization. Will be applied in this presentation as it applies to the emergence shared print and digital repositories, and their potential impact on the library service model. This framework is one that we have been exploring recently within OCLC Research as a way of understanding how libraries organize themselves – how they determine which activities to manage internally, and which ones to transact for in the open marketplace.
The ‘library service bundle’ (Dempsey, Lavoie) as it was traditionally configured has come undone as institutions selectively externalize operations and infrastructure. Changing transactions costs are certainly part of what has motivated this shift to external provisioning, but…
[click] Libraries aren’t ‘rational actors’ and potential cost savings in the marketplace are not, in themselves, a justification for externalizing core functions – like the stewardship of national heritage collections – that have cultural as well as economic value. Simply put, the incentives necessary to mobilize a widespread shift in collection management practices cannot be reduced to opportunistic economic gains. [click] Libraries will not surrender management of the assets they have accumulated unless the redistribution of resources demonstrably increases their relevance and value. And for this to happen, [click] libraries (and the institutions they serve) will need to have confidence that the agents that propose to manage the system-wide print collection – or some significant portion of it – as an aggregate resource will not seek to capitalize on its value at the expense of the library system. I am not saying that libraries will reject partnerships with commercial content and preservation partners – the academic has already effectively (and even advantageously) outsourced management of its electronic serials collections to commercial entities. But where long-term management of the global book collection is concerned , it is reasonable to expect that the library community will require a different set of services and providers. This is especially true, as the conference organizers have pointed out, because mass digitization of books in university libraries has fundamentally altered the traditional value proposition of a locally-managed print collection. [We don’t yet know if the increased discoverability of digitized books will increase or decrease use of print collections, but we can be fairly certain that libraries ]
Elinor Ostrom, a political scientist based at Indiana University, was awarded the Nobel Prize in economics earlier this month for her work on economic governance, or the study of how economic activities are managed within organizations. Ostrom’s work demonstrates that the ‘tragedy of the commons’ – the presumption that individuals will inevitably deplete and destroy resources in which they have no private ownership claim – is by no means a necessary outcome.
It may seem improbable that a model of resource management derived from empirical studies of forests, fishing grounds and Finnish reindeer herds can be extended to libraries. But in fact, Ostrom herself has argued that libraries represent a common-pool resource susceptible to the same kind of cooperative governance regimes that have ensured the long-term sustainability of wetlands, groundwater systems and wildlife. This has important ramifications for our discussion of the ‘Universal Repository Library.’ It suggests that cooperative management of the global book collection is not only feasible, but may in fact represent the best option – better than privatization, and better than government centralization. http://www.law.duke.edu/pd/papers/ostromhes.pdf
Ostrom points to Attributes of Trusted Digital Repositories (2001) as evidence of the library community’s efforts to establish ‘rules of the game’ – community-based standards for managing pooled institutional assets.
These are rough figures... NYU [~2.5M titles in WorldCat] ReCAP [~3M? titles – est. based on SRLF] Hathi [~2.8M titles in WorldCat] In an average year, NYU acquires about 100K print volumes and transfers 20K volumes to storage. The volume of storage transfers as a % of print acquisitions has increased significantly in recent years; while NYU is not yet at ‘steady state’ with regard to on-campus print collections, they are still facing extraordinary space pressure. This is the case for many if not all university libraries. And academic research institutions without any access to print storage face an even greater challenge. In the coming year, NYU anticipates weeding almost .5M volumes (or 1% of total inventory) – and they expect the pace of withdrawals to accelerate in coming years. The question they have posed is ‘ how much of what would normally be transferred to off-site storage can instead be sourced from ReCAP & Hathi ?’
Venn diagram provides a visualization of how the different parts come together. Intersections represent areas of opportunity for new service development. NYU is not contributing content to either one of these repositories; that’s what make this service model new.
http://www.sciecom.org/ojs/index.php/sciecominfo/article/viewFile/1763/1392 Recent article on efforts to develop a national repository infrastructure in Finland, notes reluctance of some universities to adopt common platform despite ‘obvious’ economic incentives. I stumbled across this article as I was preparing for this conference and was struck by the parallel between the evolution of the institutional repository movement and the repository library movement . In both cases, we observe a slow but continuous shift toward externalization of the repository function: increasingly, institutional repositories are being reconfigured as multi-institutional disciplinary repositories, with an additional layer of professional networking and identity management services. We see the center of gravity shifting from modest institutional aggregations to large-scale hubs that can more effectively sustain scholarly activity. I see this trend as continuous with the progressive externalization of operations and service that were once organized within the boundaries of individual libraries . This is the evolutionary history of the library, as it seeks to become more deeply embedded in networks of communication and exchange. As with institutional repositories, some libraries will prefer to organize and manage collections ‘internally’. But for many others externalization of the repository function will be a viable and even preferable option, IFF existing facilities can organize themselves as a robust, horizontally integrated system. The existence of large-scale shared repositories is not, in itself, enough to motivate individual institutions to cloud-source research collections, or to create the broad market for services that is needed to sustain the ‘Universal Repository Library .’ There must be a concerted effort within the international repository community to create a service framework that enables libraries to affect a massive redistribution of print collections, before the scholarly value/function of print is so changed that continued investment in these resources is no longer ‘universally’ justified.
“ Sharing resources in Finnish university libraries: re-organising the national document supply system” (2005 conference paper) We’ve measured impact on an institutional or single repository basis. NRL has shown how centralized infrastructure has enabled a progressive externalization of document supply services at university libraries. UKRR characterizes impact in terms of cost avoidance and redeployment of library resources among participating contributors JURA describes anticipated impact on cooperative relations among contributors – not the impact it might have on other institutions We need metrics for assessing the collective impact of the repository ‘system ’. And for measuring the increased impact that closer coordination can achieve.
http://dlc.dlib.indiana.edu/dlc/handle/10535/2173 There are wonderful examples of cooperative governance as it is applied to common pool resources in Finland. This paper (from the Digital Library of the Commons) describes an effort to manage wild bears and other large carnivores that would otherwise be hunted to extinction.
http://news.bbc.co.uk/earth/hi/earth_news/newsid_8314000/8314558.stm Recent proof that cooperatively managed reindeer in Finland are being picked off by cooperatively managed eagles. A reminder that cooperation doesn’t guarantee sustainability or success. EW Deming ‘survival is not mandatory’ – need to work as a system, but cooperation doesn’t confer automatic protection against selection pressures
The Gathering Cloud Constance Malpas OCLC Research Kuopio-3, 29-30 October 2009 How Shared Repositories are Transforming the Library Landscape
Overview (in pictures) Local Context Cloud Library Carnivores, large -- management of “ Entering Finnish Airspace” Photo by wili hybrid http://www.flickr.com/photos/wili/267263693
Overview (in words) <ul><li>Framework: economic governance </li></ul><ul><ul><li>organization of the library system </li></ul></ul><ul><li>Changing locus of ‘core’ library operations </li></ul><ul><ul><li>discovery, delivery, inventory management </li></ul></ul><ul><li>Repositories in the Cloud </li></ul><ul><ul><li>case study: externalizing collection management </li></ul></ul><ul><li>Implications for shared service development </li></ul><ul><ul><li>multi-institutional entities, governance models </li></ul></ul>
Discovery happens elsewhere Cultural identity in the cloud Persistence Net-workflows Addressability
"Possession means worries and luggage bags one has to drag along." Collector of stamps plants insects tickets orphans & various other things Two Views of Collection Management
Theory of the Firm (Coase, 1937) <ul><li>Organizational boundaries shift with changing transaction costs </li></ul><ul><li>Core activities are internalized to maximize economic control in uncertain marketplace </li></ul><ul><li>Operations are externalized when cost-effective alternatives emerge, enabling firm to refocus on a more distinctive service profile </li></ul>Shared print repositories embody strategic externalization of collection management
However . . . <ul><li>Changing transactions costs are not a necessary or sufficient explanation for how library system is organized </li></ul><ul><li>Externalization of print collection management activities must enable redeployment of internal resources </li></ul><ul><li>Increased reliance on shared print repositories will depend on the emergence of a new library service platform </li></ul>
Governing the Commons (Ostrom, 1990) <ul><li>Overexploitation of common-pool resources is not inevitable </li></ul><ul><li>Multi-institutional ownership of non-commercial assets is viable and may increase long-term sustainability </li></ul><ul><li>Cooperative governance can be modeled scientifically </li></ul>Can we apply the lessons of common-pool resources to repository libraries?
E. Ostrom & C. Hess Artifacts, Facilities, And Content: Information as a Common-pool Resource (2001) [Yes]
Cooperative governance of system-wide book collection will require increased coordination of shared print repositories leverage inter-institutional assets assess carrying capacity define rules of engagement
II. An Empirical Case Study: Seeding the Cloud Library
Context <ul><li>New York University Library </li></ul><ul><ul><li>5 million volumes </li></ul></ul><ul><ul><li>Limited mandate to build comprehensive local collection </li></ul></ul><ul><ul><li>Acute space pressures; major renovation in 2012 </li></ul></ul><ul><li>ReCAP – large scale print repository </li></ul><ul><ul><li>7.5 million items; low-use print books, journals, etc </li></ul></ul><ul><ul><li>Columbia, Princeton, New York Public Library </li></ul></ul><ul><li>Hathi Trust – large-scale digital repository </li></ul><ul><ul><li>4.4 million volumes; mass-digitized books </li></ul></ul><ul><ul><li>Universities of Michigan, California, Indiana, Wisconsin… </li></ul></ul>
Research Questions <ul><li>How much duplication exists between NYU’s extant holdings and the combined resources of Hathi and ReCAP? </li></ul><ul><li>What criteria can be used to model an ‘optimal’ redistribution of physical inventory? </li></ul><ul><li>What level of space or cost savings is necessary to motivate a change in NYU’s current collection management practices? Over what time horizon ? </li></ul><ul><li>Under what conditions is NYU prepared to maximize its reliance on shared print/digital collections? </li></ul>
N=5M vols Collection profiles differ Variable repository growth rates Availability / redundancy requirements vary with rights status N = 7.6 M vols ReCAP ReCAP N=4.4M vols How much is here? How rapidly will it grow?
Preliminary Findings <ul><li>~20% of NYU holdings duplicated in Hathi (so far) </li></ul><ul><ul><li>~10% duplicated in Hathi and a single print repository </li></ul></ul><ul><li>More long-tail resources than anticipated </li></ul><ul><ul><li>Hathi coverage was the surprise: >40% of titles held by <25 libraries; vs. 26% of titles at NYU </li></ul></ul><ul><li>Less public domain coverage than hoped for </li></ul><ul><ul><li>~16% of volumes (~13% of titles) in Hathi </li></ul></ul><ul><li>‘ Virtuous circle’ – predominance of in-copyright, long tail resources increases value of shared print repositories as preservation and access providers </li></ul>
Horizontal integration of existing repositories is critical to further externalization of print management activities Improved coverage Increased confidence Increased impact
externalization of repository function October 2009
Shared Repositories in the US NRLF SRLF ReCAP SORD NWORD PASCL NEORD MLAC UMO FCLD TUG 12 shared repositories in 2 decades 26M volumes in aggregate Collective impact ? Built Capacity
Measuring Impact “ In Phase I of the UKRR, 8 libraries repurposed 11,000 metres of shelf-space , representing reduced estate costs of £308K …” Anticipated impact of JURA includes “ fostering cooperative collection development between 8 institutions” and “ last copy storage of printed journals for consortia ” We need metrics for assessing collective impact of repository system P. Vattulainen “Sharing resources in Finnish university libraries: re-organising the national document supply system” (2005)
Some Recommendations <ul><li>Characterize, quantify value that a globally coordinated repository platform will create </li></ul><ul><ul><li> articulate a shared service profile </li></ul></ul><ul><li>Communicate distinctive value of shared infrastructure </li></ul><ul><ul><li> advocate for increased institutional reliance </li></ul></ul><ul><li>Cultivate inter-institutional governance models to ensure long-term sustainability </li></ul><ul><ul><li> accelerate horizontal integration </li></ul></ul>
Practically speaking . . . <ul><li>Maximize network disclosure of repository collections and services </li></ul><ul><li>Develop a business plan for non-content contributors </li></ul><ul><li>Scrutinize duplication between local repository holdings and mass digitized corpus </li></ul><ul><li>Acknowledge (and address) risks that shared print model poses for traditional paradigm of institutional autonomy </li></ul>