Close this search box.
Close this search box.

Collections: Library Collections Online


by Abby Smith, Council on Library and Information Resources

A cursory glance at a handful of library and museum Web sites reveals that these two kinds of institutions conceptualize their Web presences quite differently. Each of their core missions, though being in a broad sense about education, mandate different collections and, in turn, different service of those collections. Rather than discuss library and museum collections as a whole, therefore, Bernard Reilly and I will address them separately, even at the risk of emphasizing the distinctive rather than the common features of their Web sites. In the end, museum objects and library items are indistinguishable from one another when transferred to digital form. A digital Blue Boy and a digital Huckleberry Finn share the same behaviors, demand the same creation of metadata and management tools, reside on the same network, and are retrieved onto the same computer screen. Does this encourage virtual visitors to blur the distinctions between these entities in ways they could not if they were actually visiting a reading room or gallery? If so, does it make a difference?

The world of Huck Finn is one of ever-receding horizons and serendipitous finds. The general expectation of library visitors is that a library is like the Mississippi: all of life is there-whatever they want they can find-and it will cost them nothing to access but a little of their time and a very good river pilot. In reality, few library collections are as easy to navigate on the Web as they are onsite, but the exceptions appear to have learned a lot from museums and have focused on mounting highly curated objects of cultural interest rather than large, unedited collections.


Whereas art museums tend to collect the rare or unique, research libraries are more like natural history museums. They build collections that provide a comprehensive source base for researchers to use onsite. In many ways and, again, in contrast to art museums, libraries’ collections must be redundant of other library collections in order to make things readily accessible to their patrons. Like natural history museums, they tend to have collection strengths in specific subjects, and, within those, frequently have scarce or unique materials of high artifactual value. Library special collections that contain unpublished and nonprint sources often contain items that are rare or unique. Taken item by item, they may be of minimal financial value and may have research value only to the extent that they are part of a larger whole.

Libraries value comprehensiveness because the only way to make a rich resource base available to researchers is to have as much onsite as possible. The availability of digital resources, with their promise of instantaneous access to information in remote locations, means that libraries do not have to own an item in order to serve it to their patrons. Redundancy of resources is not valuable in the digital library, where information is independent of any physical medium and access can be provided without proximity to the item. In fact, redundancy should be seen as a waste of resources. This limit on redundancy has serious implications for collection development-at least for collections of digital information-but it should also influence what analog materials are converted to digital form and how.


The primary service that libraries offer is access to their collections-access to information about their collections and to the information contained in them. Unlike museums, libraries do not define access to include interpretation; on the contrary, the rawer the materials served, especially in a research library, the better. The more “cooked”-that is, selected, edited, shaped by an expert-the less integrity an item is deemed to have as an object of research. Librarians are not curators; they are not expected to have deep substantive knowledge about their collections, to provide historical and contextual interpretation, or to make judgments about objects. A subject specialist in a library should, rather, be expert in the source base of one or more domains of information to build an excellent collection that can be used and interpreted by the researcher. The librarian’s responsibility is to acquire the best resources, organize them for ready access, and preserve them for future use.

What happens then when the real becomes virtual? How does digital transformation affect the library item, and how does that affect the services that libraries offer or, more precisely, the terms of service?

Digital Collections

Digitization of analog materials is, despite its revolutionary nature, usually treated as just a service-a service that provides new forms of access to analog materials. Digitization is essentially a superior form of copying-one without loss or, as we whimsically say, “lossless.” It defies the physical constraints of time and distance and provides the chance to look at many disparate collections at a time and place convenient to the researcher. It can also significantly enhance information retrieval from small, damaged, or poorly preserved items through various types of image manipulation. Given the as yet unsolved problem of digital longevity, digitization is not accepted as a form of preservation reformatting, as microfilming is. Digitization serves preservation goals only to the extent that a digitized copy of an item can be served in lieu of an original and, hence, can reduce the stresses of physical handling-not an inconsiderable boon to rare book collections and photo archives. Nevertheless, there are cases where the availability of an item through a digital surrogate actually increases demand for an original because digitization is, it turns out, also a form of advertising.

Among the unintended consequences of digitization is the creation of additional collection items-that is, we now have more items in the collections to care for. A digitized item is a surrogate but more than a surrogate, a copy but more than a copy. Because it transforms an analog item into a digital one, digitization fundamentally alters the way that the original object holds and conveys meaning.

A semiotician might say that a digital version of, for example, a first edition of William Blake’s Songs of Innocence signifies in a different way from the original tome and, therefore, it signifies a different thing. The content of the message has changed because the medium has changed. Although this might strike some as arcane to the point of irrelevance, at both deep and superficial levels, the analog and digital forms of recording information are structurally and functionally radically different. Therefore, a digital copy is a radically different thing than the analog text or image or sound from which it was created. Given that the digital version is expensive to produce and also has a great deal of value added, the institution that is in the business of extending access through digital imaging should also be in the business of protecting and preserving its new, digital asset.

More than museums, libraries will view the digital collection item as one that should be or has been permanently accessioned into their collections. The Library of Congress does not plan to rotate items in and out of its National Digital Library. The items on the Digital Schomburg site at the New York Public Library constitute not an exhibition, but a collection. The site is an information resource that will, in all likelihood, not be deaccessioned. And, as Anne Kenney points out, the existence of digital back issues of scholarly journals on JSTOR has not led to deaccessioning the physical copies. Where is the cost saving?

Digital Services

Digitization changes the nature of service to patrons. Service of collections across the Internet is different from service in the reading room in ways that both enhance and undermine the traditional ways of conducting research with primary source materials. The Blake site at the University of Virginia1 not only allows one to spend time with a virtual version of Songs of Innocence but also to compare several different versions of the same text held at different locations. Books that have been digitized to allow full text searching are transformed from a simple hard-copy text into a database that can be queried in ways that were formerly impossible. This easy access can be extraordinarily addictive, and it is no wonder that digital patrons are called “users.”


Not all information resources in a library’s collection are good candidates for digital conversion. On the contrary, only a rather small proportion of items has potential to be digitized effectively. The following list describes some real-life constraints.

  • Technology. The creation of digital surrogates is ideal for items that are rare or fragile. It is also ideal for items that share similar provenance or that in some way are deeply connected but are housed at different locations. On the other hand, computers lead to a bias toward text over image and toward still image over moving image or sound. Oversized materials, such as architectural drawings, and materials in poor physical condition also are inimical to digital capture at this time.
  • Intellectual Property. Because of copyright and its discontents, we find a bias toward selecting materials for digitization that are in the public domain. This bias is a great boon to Victorian and early modern materials, not to mention to fans of Beowulf or Thomas Jefferson. But the idea that a library’s information resources can be adequately or meaningfully represented by a preponderance of public domain materials is senseless. Libraries are committed to continuing the traditional service that they have been providing to patrons in the actual world, where information is always fixed to a medium. This service has been free, and, in all but exceptional circumstances, materials that are both in and not in the public domain are equally accessible. This tradition is not practiced online. It is a cause for great concern among most professionals, and it should be a cause for equal concern among us all.
  • Resources. Digitization costs a great deal of time and money. The presumption that libraries can afford to serve their collections digitally in addition to serving them onsite or through traditional interlibrary loan must be questioned. How can libraries afford the digital services that they wish to provide? We face a risk that, in order to subsidize the large financial outlays that digitization demands, libraries will be tempted to extract profits from digitization of their “museum” pieces, that is, items such as rare manuscripts or books that are of high artifactual value. Library collections are assets, but they seldom have been income-producing assets. Poor Huck, casting a sidelong glance at Blue Boy in the museum, sees that digitization just might offer the opportunity to cash in on his library’s moiety of what has recently been dubbed “cultural heritage.”

Some Consequences

The Marketing of Collections

Treating information resources as cultural heritage that can be marketed is a new idea for libraries. How would such marketing work? If a library can charge for some things, should it charge for everything, just to keep its franchise intact and protect the brand name? Consider, for example, a private university that has extraordinary holdings of rare books-books that are available to undergraduates for use in course work and that are deemed a great resource for scholars worldwide. If these texts were digitized, they would add a significant asset to the global digital library. But would they at the same time diminish, even if slightly, the allure of an undergraduate degree from this university by making what could only have been purchased with tuition-access to these books-available to everyone with an Internet connection? No one has quantified the relationship between a library’s collections and services and the price of an undergraduate or graduate degree; however, few potential students who visit a college or university have not been shown the library and told of the richness of its collections and the advantages of access to these collections.

The Role of Specialists

The process of selecting items for digitization requires librarians to assume new and often uncomfortable roles. They are forced to play editor, selecting parts of large collections to put online rather than serving up the whole and letting the user select. They are also asked to censor items in a collection that, although appropriate for service in a reading room, are not appropriate to be broadcast across the Internet. Librarians do not like to censor research collections. The need to create metadata has the potential to turn catalogers into curators, for creating metadata involves creating a context that provides layers of information to facilitate retrieval and interpretations.

The Role of the Institution

One of the things that makes most library sites less appealing than museum sites is the paucity of interpretive materials. Digital visitors want information that has been assimilated. The Library of Congress’s American Memory site is successful, in part, because of the way it is curated. Collections that would take days to track down physically on Capitol Hill are brought together on the Library of Congress home page and grouped coherently and thematically. This presentation masks the fact that, in real life, these collections are not ordered, described, or served in relation to each other. Sites that reflect more closely the actual physical arrangement of collections, such as those of Yale or the University of Michigan, leave the online visitor, like the onsite visitor, to go from one library collection to another, without a central portal. This makes the visitor feel like he or she is traveling the Mississippi without a pilot. Museum sites, by contrast, direct visitors to their digital collections and provide more information. Moving from the wide, expansive, barely mapped world of Huck Finn to that of the Blue Boy, we find the surrounding landscape tamer, the green, open spaces well manicured, and the paths better trod.


1 Available from

Skip to content