What is a library anymore, anyway?
First Monday

What is a library anymore, anyway? by Michael A. Keller, Victoria A. Reich, and Andrew C. Herkovic

What is a library anymore, anyway? by Michael A. Keller, Victoria A. Reich, and Andrew C. Herkovic
Libraries in the future will undertake local control, especially for long-term preservation and accessibility of digital as well as analog collections. Failure to embrace that role would cause libraries and librarians rapidly to lose relevance and value as Internet and other digital resources develop. Local control of collections is critical both to assure permanence and to provide a key degree of selectivity, which — contrary to the irrational exuberance of making everything available to everybody — is vital to providing service to communities of readers. Librarians need new tools, such as the LOCKSS system, to enable both persistence and selection of electronic information.







We have observed a propensity for information technologists to predict with complete confidence the imminent demise of libraries. The seeds of this prognostication may date back to Vannevar Bush's seminal paper of 1945 [1], but the forest of such predictions has grown thick in the past decade. In our observation, the confidence with which such predictions are made is inversely proportional to the predictor's professional habitual use of published information. That is, the prediction that libraries are becoming obsolete or useless is a projection onto the world of the internal set of the speaker and reflects a lack of appreciation of libraries' deep, often hidden functions, especially in the realm of digital information resources. Long-term or intensive library users, on the other hand, rarely if ever, in our experience, foresee this demise, except as a distopian nightmare. And indeed, we often note that when some technologists talk of information, it is as an essentially quantitative abstraction, something to manage rather than use — in short, a commodity — not as the lifeblood and substance of scholarly [2] inquiry and endeavor.

So to some extent the question of what is a library anymore could be one about which, "if you have to ask the question, you wouldn't understand the answer." Our intention is to attempt the answer anyway. This intention is inspired, if not modeled, on the remarks of Gerhard Casper, then president of Stanford, at the 1999 rededication of the restored central campus library, "Who Needs a Library, Anyway?" (Casper, 1999)




As librarians, our task has always been and remains to use the tools of information technology to serve the needs of information workers or seekers, bearing in mind that all recorded information and its support infrastructure — books and shelves no less than online search engines — is information technology. Actually, understanding what is a librarian anymore may be easier than understanding what is a library. There can be, and increasingly may be, librarians without libraries (in the sense that they are not based at a single physical institution). But can there be libraries (or even the Universal Virtual Library as some have imagined) without librarians? Or without tangible institutional existence (per the bricks and mortar cliché)? Obviously, many think so. Others, particularly those responsible for institutional budgets, may hope so. We think this is nonsense.

Ignoring the physical, technological underpinnings for now, we assert that the library is, at root, a collection of information selected for use of, and made useable for, a particular community [3]. That community may be large or small, physically proximate or not, present or future, homogeneous or not, but it is essential that it be identified and at least partially understood. That is, proverbially like politics, all collections are local.

The roles of collections in library services were enumerated by Michael Buckland in his 1992 manifesto [4] as preservation, dispensing, bibliographic and symbolic. He added, "If these are the four purposes of collections of library materials, we need to inquire how the change from paper to electronic media may change how we seek to effect these roles." He went on to discuss the "'owning versus borrowing' trade-off" in the context of electronic vs. paper-based documents, while acknowledging that use, law, and publishing were in a state of flux. Since 1992, nearly everything about electronic information has changed, but we think many or all of the fundamental services and roles to which Buckland refers remain in principle unchanged. Our own enumeration of these functions is as follows: Selection and acquisition for collections; provision of intellectual access to local and remote collections; interpretation and discovery information for patrons; distribution of information resources; and preservation of those resources. To these five, which obtain equally well for digital and analog resources, we may add a sixth, the provision of clean, well-lit, relatively quiet places for books and readers, a physical function with some implications for virtual library spaces.

No collection is perfect or complete. Thus, there is a legacy, dating back at least for a millennium and a half, of interlibrary loan. We owe the existence of almost all ancient texts, with the exception of those preserved in cuneiform, to the practice of individual libraries or scriptoria borrowing and making local copies for local use of other libraries' manuscripts. They did not do so to save Western Civilization; they did so to have and to hold those texts for local purposes. So, today, if a scholar at Berkeley, to use a parochial example, needs a book not available at the University of California, she may use Stanford's copy either by visiting or simply by requesting it online. In an imagined future, she might expect the complete work to be available electronically and instantly. And while we are deeply engaged in creating electronic versions of existing texts and devising ways of making them more useful to readers and scholars, we know there are profound issues about the bland assumption that everything will just be there online someday.

The extension of this assumption is that, once content ceases to be a local concern, there will not be a need for Berkeley, or Stanford, or anyone else to hold collections. This brings us back to the notion that libraries are obsolete and fated for oblivion. For some years, some in the library field have been creeping in this direction under the rubric "access, not ownership." There is a trend toward redefining what constitutes a collection:

"Instead of describing collections as "those things owned", a better definition may be "information resources for which the library invests financial resources — directly or indirectly — to manage, service, or preserve on behalf of library users, regardless of the location of content." "Collections" now include resources owned by the library and those accessed in remote locations; the norm is now an interdependent mix of ownership and access, with the location of the material increasingly irrelevant to users" [5].

As recently stated by the ARL Collections and Access Issues Task Force, librarians are "Expanding the Definition of Collections":

"Libraries have expanded the traditional view and definition of collections so that the concept no longer equates with those materials that the library "owns". The boundaries have expanded far beyond the print collections on site or the electronic files mounted locally to include electronic materials licensed or managed by the library and materials available through consortia. Increasingly libraries are taking responsibility for born-digital collections (such as geo-spatial or numeric data sets, faculty or class Web sites) and developing tools for their management and use. In a growing number of cases, a library's collection also includes resources that reside outside the domain of the library but for which the library takes some responsibility for managing and servicing" (emphasis added) [6].

The centrality of access is not at issue here. Our concern is not with the library's proper role in providing such collections, nor is it with providing tools (read, Web pages) for their use. We are concerned, however, that this masks an evolution from a model of professional guidance to a community of readers — which we heartily endorse and practice — into a larger vision of "libraries" abandoning responsibility for physical collections — a dysfunctional and possibly doomed vision, in our view. A Web page with a set of links is not a library. If we understand libraries simply to be nodes in a global digital network of common resources, nobody retains any responsibility for either those materials or serving the needs of the community. In this environment, there would indeed be no meaningful role for libraries. We do not see this as serving the needs or interests of our readers, particularly if viewed over time. We have come to understand "access, not ownership" more as rationalization of constrained choices than as a functional understanding of libraries' service to their communities, other than purchasing and licensing agents.

It has been true for several years that the printed editions of some leading scientific journals are no longer complete or authoritative; the authentic versions of record are the electronic editions, which may have more articles, embedded or linked supplementary information (in electronic form) appended to the articles, or editorial content that cannot exist in print form (simulations, animated models, video or sound recordings, etc.). This is a wonderful development [7], and for this among many reasons we favor subscribing to the electronic edition, but it leads to the concern that we have no way to assure that this material will be available to our readers in one, two, or ten years' time (whether or not we attempt to continue subscribing to the journal). Whether we keep the printed edition or throw it away, our readers' future access to the full journal contents is in jeopardy.

At a meeting in Philadelphia in January 2003 (in conjunction with the American Library Association's midwinter conference), Theodore Bergstrom reportedly "got librarians' attention by asking a provocative question: just why are libraries involved in subscribing to e-journal site licenses when the e-journals aren't residing in the library, and are being used largely outside of the library?" [8] Subsequent discussion raised the question of boycotting site licenses, apparently for the purpose of redressing for-profit publishers' predatory pricing and bundling policies. Without commenting on the propriety of that strategy, we suggest a corollary motivation: The impermanence of what site licenses make momentarily accessible [9].

Somebody somewhere will do something, right? Lots of things have been tried, at least in principle, most of which have depended on: Lead agencies acting ostensibly on behalf of all others indefinitely; and, publishers abdicating control of their proprietary content to those lead agencies.

Despite the best of intentions, such approaches have been largely stillborn. We are willing to assume that such approaches will eventually take hold at least selectively and indeed, we intend to take part in them. However, they will likely be highly selective, as to both content and beneficiary, and will be gambles on the part of librarians and leaps of faith on the part of publishers. Meanwhile, the vast majority of libraries are left as passive holders of others' promises. We think some professional skepticism is called for.

Publishers, may be inclined at times to "un-publish" articles and, in the case of the U.S. Government, even entire volumes [10]. If they control the only server on which electronic material exists, there is no recourse, a grotesque disservice to science, to scholarship, and to the public good. Our colleague Jim O'Donnell wrote the following to the Liblicense-L discussion:

"The discussion over publisher-removed articles is of course a discussion over the reliability of archives. We are accustomed to being able to go back to published material long after the fact and to find a stable and accurate record of what was said. Traditionally, libraries have been the guarantors of this process: Preserving many copies, with no legal liability for the content (or at least less than the publisher might have), and with an institutional commitment to permanence and preservation. The "vanishing act" discussion highlights a feature of unreliability of e-archives that depends (1) on the physical malleability of the record and (2) on the slightly lower commitment to full preservation that a publisher might have. It is disturbing, because it is the tip of the iceberg, I think: If for fairly transient reasons, publishers will pull articles, when might not publishers prove unreliable for other reasons?

But the question that follows on this discussion for me is this: If we were to ask that not publishers but authors be the guarantors of permanence, self-publishing or publishing in institutional repositories where the author retains control over the copyright and disposition of his/her material — what protection do we then have to assure us that articles will remain archived, unchanged, in perpetuity? Are there articles I have written that I wouldn't mind disappearing? Actually, yes. Are there pieces of articles that I would quietly change if I could? Well, interesting thought, sure.

Is it important that the record abide? Then should not all discussions of e-publishing for scholarly purposes include a discussion of preservation that includes not only the physical vulnerability of the media but their psychosocial vulnerability? What guarantors other than libraries do we realistically have?" [11]

Some publishers make one or another surrogate available to subscribers on tape or CD-ROM, with or without charge. This creates considerable challenges for maintenance and management, and may or may not feature the links and other functions that make the online edition distinctively useful. For example, Reed-Elsevier makes its journal content available to subscribing libraries on tape. Few libraries [12] in North American are known to load the Elsevier tapes, presumably because of both cost and storage management issues. So, almost everyone depends on continued vendor-based access to these materials. Tape delivery is not even an option for most publishers' journals or databases (or any other type of material). Is this an acceptable situation for libraries? Instinctively, we think not. As we stated earlier, a Web page with a set of links (to publisher sites) is not a library, and a Web page with a set of obsolete, denied or expired links is nothing at all.

For these reasons, Stanford is developing [13] the LOCKSS system and protocol, which allows libraries to create, manage, and maintain persistent caches of e-journal content to which it subscribes. We will not here go into detail of how LOCKSS works [14], but its salient features are that each library that wishes to physically possess and control e-journal content to which it has subscribed can do so, with minimal effort and cost, on behalf of its readers. When a publisher's Web site goes down — be it for an hour or forever — the library can serve up that content to its users. The library is no longer reduced to being a Web page with a set of links — a mere pass-through — vis-à-vis e-journals. The traditional virtue of libraries — to assure access to a community of readers by holding selected content of importance to those readers — is reasserted. Another virtue of libraries, that they have a high degree of collective redundancy, is also replicated in the e-journal sphere by the LOCKSS approach of multiple caches at different institutions. We emphasize that redundancy is a virtue, as structural engineers have learned. Any system or environment that rests on a single point of failure, whether it be the Alexandrine Library (ver. 1.X), a hard disk, a vendor's Web site, or even the Library of Congress, is subject to failure.

Our colleague Harold Billings at the University of Texas Austin raises these issues clearly:

"I have just suggested to my staff that we should "add" these publications [two guides to open-access publishing released online by the Budapest Open Access Initiative] to our collections, and "catalog" them. But without exploring them too deeply, I then immediately wonder: How do we know these publications, digital in PDF, will continue to exist after we add and catalog them? This is obviously the same kind of question that one has to ask of all electronic publications, journal or otherwise. The most secure method would be to grab them and download them into our own servers. I guess this illustrates how important the ongoing existence of any publisher is. This is why I find individual archiving to be without any merit, while institutional archiving — and subsequent networking — could have an important future if the archive is maintained by a trusted university, organization or commercial entity. This is where the concept of Stanford's LOCKSS comes into important play [15].

This is even more significant in the case of electronic-only journals, i.e., those that do not exist at all in print (and may be sitting unprotected on somebody's desktop server). E-journals constitute a single, albeit vital and costly, class of collection materials. There is simply no way to collect the rapidly growing corpora of lost-cost or no-cost Web-published materials which are of incalculable value and interest, but that are unlikely to survive the years, to say nothing of the decades. Web sites have taken on the historical roles and research value of samizdat, avant-garde magazines, seditious literature, fringe political manifesti, etc. The Internet houses the New Underground: highly specialized, highly controversial, sometimes dangerously political, and all extremely ephemeral. In reference to literary and critical journals not being collected (because they exist only online), our colleague William McPheron states, "If they were in paper, we'd be getting them." At present, he cannot follow his curatorial instinct to collect and retain them.

Several libraries, working with the Center for Research Libraries, are currently conducting studies regarding the collection and retention of political Web sites in various parts of the world; without such efforts, much of contemporary political evolution in unstable regions may be permanently untraceable, precisely because 1) the documents of interest were published only on the Internet, and 2) nobody — with the possible and notorious exception of the winners — will make the effort to collect and retain them [16].

It is worth noting that a group of American government documents librarians are exploring the possibility of adapting the LOCKSS system to build and maintain collections of U.S. and local government documents [17]. Others are exploring applying LOCKSS to medical gray literature (mostly evidence-based) and other corpora of literature, such as university theses and dissertations.

What we (as interlocutors for our readers) need is not indiscriminate warehousing of ephemeral Web sites, but selective, intelligent, targeted collection development — exactly what librarians have always done and, by and large, done well in service to their local communities. In clear distinction to the prevailing dynamics of the Internet — "Fast, Cheap, and Out of Control," to borrow a documentary film title [18] — we need to be deliberate about what we gather, control carefully what we do gather, and even more deliberate about discarding information [19].

New tools are needed to find, evaluate, select, and preserve such content. Whether they will resemble LOCKSS, how labor intensive they will be, how affordable they will be are all unknown. However, we assert that such tools will provide value and utility only to the extent they allow individual libraries to make their own collection decisions, to control file content locally, and to serve their own communities through their own collections.




Whether or not consciously, libraries and librarians have long been prominent among the few kinds of social agencies that have preserved continuity of cultural heritages. By serving as custodians of local collections, they have incidentally served a larger common good. Whatever other public benefits they provide, publishers and Internet promoters do not, and cannot be expected to, fulfill this custodianship role.

That libraries may be becoming obsolete is, to some degree, plausible, not because they are losing some kind of competition with the Internet for eyeballs or compellingly superior content, but rather because libraries may be in the process of abandoning their role as collection builders and managers. We do not suggest that anybody is to blame in this as a practical matter; we haven't yet figured out how to retain that role in an increasingly (if never exclusively) digital information universe. The economics and limitations of technology (with some enormous intellectual property issues on the side) are serious, if temporary, impedimenta. But if we don't assert the importance of that role, of the centrality of selecting, acquiring, retaining, preserving and building means of access to collections, we will inevitably fade away. End of article


About the Authors

At Stanford, Michael A. Keller is the Ida M. Green University Librarian, Director of Academic Information Resources, Publisher of HighWire Press, and Publisher of the Stanford University Press. These titles touch on his major professional preoccupations: Commitment to support of research, teaching and learning; effective deployment of information technology hand-in-hand with materials; active involvement in the evolution and growth of scholarly communication. He may be best known at present for his distinctively entrepreneurial style of librarianship. As University Librarian, he endeavors to champion deep collecting of traditional library materials (especially of manuscript and archival materials) concurrent with full engagement in emerging information technologies.

Keller was educated at Hamilton College (B.A. Biology, Music 1967), SUNY Buffalo (M.A., Musicology, 1970), SUNY Geneseo (M.L.S., 1971), and SUNY Buffalo (a.b.d. Ph.D., Musicology). From 1973 to 1981, he served as Music Librarian and Senior Lecturer in Musicology at Cornell University and then in a similar capacity at UC Berkeley. While at Berkeley, he also taught musicology at Stanford University and began the complete revision of the definitive Music Research and Reference Materials, an annotated bibliography popularly known as "Duckles" in honor of its original compiler. Yale called him to the post of Associate University Librarian and Director of Collection Development in 1986. In 1993, he the Ida M. Green Director of Libraries at Stanford. In 1994, he was named to his current position of University Librarian and Director of Academic Information Resources. In 1995, by establishing HighWire Press, he became its publisher, and in April 2000, he was assigned similar strategic duty for the Stanford University Press.

Vicky Reich works to facilitate the industry's transition from print to online publishing models. She is Director and co-founder of the LOCKSS Program, which allows libraries to retain local collection control of materials delivered through the Web while preserving the functionality of the original Web-based content (see http://lockss.stanford.edu). Prior to the LOCKSS Program, she was, for eight years, the Assistant Director of HighWire Press, during which she led its Market Support Services to build international relationships with academic and corporate librarians. She has over 20 years of extensive library experience in both public and technical services and in both public and private institutions. She has held positions at the Upjohn Company; University of Michigan; Library of Congress; National Agricultural Library; and, Stanford University. She earned her MLS from the University of Michigan.
E-mail: vreich@stanford.edu

Andrew Herkovic is responsible for foundation, corporate, and government relations at the Stanford University Libraries. His duties involve public relations, project management, speech and proposal writing, publishing, and liaison with partners in library initiatives. Past project interests include investigation of the scholarly use of e-journals; perceived crises in scholarly communications based on journal pricing and rights management; and, leadership development for librarians. His has held positions in academia and business since his first professional incarnation as a library cataloger at Cornell University, of which he is a graduate.



The authors thank Don Waters of the Andrew W. Mellon Foundation; the National Science Foundation; and, Sun Microsystems for their continued support of the LOCKSS Program. We also thank the many libraries and publishers who are participating in the beta test and who are giving of their time and resources to make this program viable.

Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.



1. Vannevar Bush, 1945. "As We May Think," Atlantic Monthly, volume 176, number 1 (July), pp. 101-108; and, at http://www.theatlantic.com/unbound/flashbks/computer/bushf.htm.

2. Scholarly in the inclusive sense of informal as well as formal inquiry.

3. Although a good library is as much a service organization as a collection, we focus for now on the collection aspect.

4. "Chapter 6: Collections Reconsidered," In: Michael Buckland, 1992. Redesigning Library Services: A Manifesto. Chicago: American Library Association; Internet edition 1997 at http://sunsite.berkeley.edu/Literature/Library/Redesigning/html.html.

5. "Collections & Access for the 21st-Century Scholar: Changing Roles of Research Libraries, A Report from the ARL Collections & Access Issues Task Force," ARL Bimonthly Report 225, December 2002 http://www.arl.org/newsltr/225/main.html.

6. Ibid.

7. Encouraged and implemented by Stanford's HighWire Press, among others.

8. "Economics Lesson Leads to Inspiration: Is a Site License Boycott by Libraries Possible?" Library Journal Academic News Wire, 30 January 2003.

9. This begs the obvious question of what demand might such an action invoke, and we think there are several possible answers, ranging from the very vague, e.g., some tangible guarantee that subscribers at least will never lose access to content for which a site license was ever bought, to the very specific, e.g., active cooperation with the LOCKSS system.

10. The reader may wish to review the recent exchanges about Elsevier's removal of articles on the Liblicense-L discussion list archive at http://www.library.yale.edu/~llicense/index.shtml.

11. E-mail from Jim O'Donnell to Liblicense-L Wed. 29 January 2003, Subject "Re: vanishing act", at http://www.library.yale.edu/~llicense/ListArchives/0301/msg00118.html.

12. The Los Alamos National Laboratory, University of Toronto, Ohio Link, and National Library of the Netherlands are archiving this material as well.

13. With generous grants from the Andrew W. Mellon Foundation, National Science Foundation, and Sun Microsystems.

14. Which is available at http://lockss.stanford.edu.

15. E-mail From Harold Billings to fos-forum@topica.com Thursday, 30 January 2003, subject "Re: two important guides to open-access publishing".

16. "Political Communications Web Archiving: A Proposal to the Andrew W. Mellon Foundation," Center for Research Libraries, 26 July 2002. The Mellon Foundation has funded this project.

17. Funded by a Small Grant for Exploratory Research from the National Science Foundation.

18. http://www.sonypictures.com/classics/fastcheap/.

19. By no coincidence, the logo for the LOCKSS system — visible at http://lockss.stanford.edu/ — features a tortoise as a metaphor for being slow, but long-lived. This metaphor appears in a number of cultural contexts, ranging from Aesop to Native American legend and art.



Association of Research Libraries (ARL) Collections & Access Issues Task Force, 2002, "Collections & Access for the 21st-Century Scholar: Changing Roles of Research Libraries, A Report," ARL Bimonthly Report 225 (December), at http://www.arl.org/newsltr/225/main.html.

H. Billings, 2003. E-mail to fox-forum@topica.com (Thursday, 30 January), Subject "Re: two important guides to open-access publishing".

M. Buckland, 1992. Redesigning Library Services: A Manifesto. Chicago: American Library Association; Internet edition 1997 at http://sunsite.berkeley.edu/Literature/Library/Redesigning/html.html.

V. Bush, 1945. "As We May Think," Atlantic Monthly, volume 176, number 1 (July), pp. 101-108; and, at http://www.theatlantic.com/unbound/flashbks/computer/bushf.htm.

Center for Research Libraries (CRL), 2002. "Political Communications Web Archiving: A Proposal to the Andrew W. Mellon Foundation," (26 July), at http://www.library.cornell.edu/iris/research/WebPolCom.pdf.

G. Caspar 1999. "Who Needs a Library, Anyway?" [Text of President Gerhard Casper's remarks to the Stanford community at the dedication of the Bing Wing of the Cecil H. Green Library on 12 October 1999], at http://www.stanford.edu/dept/news/report/news/october13/libtext-1013.html.

Library Journal, 2003. "Economics Lesson Leads to Inspiration: Is a Site License Boycott by Libraries Possible?" Library Journal Academic News Wire (30 January).

J. O'Donnell, 2003. E-mail to Liblicense-L (Wednesday 29 January), Subject "Re: vanishing act", at http://www.library.yale.edu/~llicense/ListArchives/0301/msg00118.html.

Editorial history

Paper received 13 February 2003; revised 19 February 2003; accepted 12 April 2003.

Contents Index

Copyright ©2003, First Monday

Copyright ©2003, Michael A. Keller

Copyright ©2003, Victoria A. Reich

Copyright ©2003, Andrew C. Herkovic

What is a library anymore, anyway? by Michael A. Keller, Victoria A. Reich, and Andrew C. Herkovic
First Monday, volume 8, number 5 (May 2003),
URL: http://firstmonday.org/issues/issue8_5/keller/index.html

A Great Cities Initiative of the University of Illinois at Chicago University Library.

© First Monday, 1995-2019. ISSN 1396-0466.