Technological changes challenge the raison d'¨ºtre of research libraries.

Stephen Phillips meets a man trying to ensure their long long-term future

Brian Eno has enjoyed success at the cutting edge of popular music and in other artistic ventures. Hardly, it would seem, the sort to keep company with librarians, who are not generally credited with great creative vision. But just as Eno is no ordinary musician, Michael Keller is no ordinary librarian. The chief librarian of Stanford University, who corresponds with Eno by e-mail, is something of a maverick to his colleagues.

Keller, a keynote speaker at next week's Society of College, National and University Librarians conference in Newcastle, has a PhD in musicology. He had served as music librarian at Cornell University and at the University of California, Berkeley, and then as Yale University's associate university librarian before he joined Stanford in 1994. His correspondence with Eno has helped him to think more creatively about his role as a "cultural custodian", he says.

É«ºÐÖ±²¥

He and Eno sit on the board of the Long Now Foundation, a San Francisco think-tank and self-styled "very long-term cultural institution" that was conceived in 01996 (the extra digit is to foil the "deca-millennium bug"

looming just 8,000 years ahead, the website explains) to counter short-termism in society and to foster visionary thinking.

É«ºÐÖ±²¥

The group's best-known project is the Clock of the Long Now, which is intended to "embody deep time" by ticking for 10,000 years in a cave in the Nevada desert. But its work also holds particular relevance for research libraries. Its offshoots include the Rosetta Project, which is dedicated to preserving, on disk and in online collections, material documenting the languages of "small, unique, localised human societies" that are threatened by globalisation. There are also various "digital continuity software projects" aimed at preserving electronic documents for posterity.

In 2000, Keller hosted the 10,000 Year Library conference at Stanford, at which delegates pondered issues expected to confront the profession over the ultra-long haul. "It's an exaggerated statement of the problem set," he concedes, noting that 10,000 years outstrips the antiquity of the oldest surviving comprehensible human documents - cuneiform tablets from present-day Iraq dated to 5,000 to 7,000 years ago - but it "forces thinking (farther out), not just managing into the next decade".

More immediately, librarians are wrestling with the implications of the web and other technological developments. Fundamental questions about the role of libraries have been raised by the proliferation of digital content and commercial text-scanning initiatives such as that by Google, which aims to make searchable over the web at least 15 million tomes from Stanford, Harvard University, the University of Michigan, Oxford University and the New York Public Library. "There's an implied debate about whether libraries are needed at all," Keller notes. "What are libraries to do? Give up and become museums of what they have on their shelves at the moment?"

In an essay written in 2003 for Washington DC's Council on Library and Information Services, Keller observed: "There is much confusion about whether the research library of the future will be some magnificent virtual collection of sources and services, owned by no one but vital to allI, whether the information age and the internet will dissolve the traditional hierarchy of significance based on the size and sophistication of libraries, or whether both of these transformations will be affected."

Efforts to get to grips with the changing information landscape are proceeding on several fronts. There is the National Digital Information Infrastructure and Preservation Program led by the US Library of Congress, which aims to hammer out the most "viable and sustainable options for long-term preservation" of digital content, and there are commercial projects such as Microsoft's collaboration with the British Library to digitise 100,000 of its books. Keller says it is unclear which approaches to "digitising archives" will prevail. "There is a lot going on, (but) we do not have enough experience with experiments to have a taxonomy of approaches."

Still, there are some guiding principles. "Redundancy counts, and large-scale redundancy counts (even) more," Keller says. Since 1999, Stanford has spearheaded development of "Lockss", which stands for "Lots of Copies Keeps Stuff Safe", online archiving software that maintains multiple copies of electronic journals in case content becomes unavailable over the web. It is now used by 80 libraries globally.

The ephemeral nature of digital documents is a pressing problem for libraries. Keller asks: "How does one think about curatorial responsibility in the light of documents that are difficult to preserve and might be visible for (only) short periods of time (many US government documents are posted online then removed within 100 days)?"

É«ºÐÖ±²¥

Extending the redundancy principle, Keller is wary of information concentrating under any single entity, especially in the light of risks such as power cuts and terrorist attacks. Another concern is accountability. "You do not want to leave the task of recording civilisation to commercial entities. Who knows whether Google would let stuff drop out," he says.

É«ºÐÖ±²¥

Despite such worries, Stanford is an enthusiastic participant in Google's book-scanning project. In a recent presentation to Stanford officials, Keller notes that it promises to "digitise every book in (the) library with no damage to the book; return... a digital copy for preservation and other purposes; (and) present... the world with a combined word index to millions of books".

As of mid-May, Google was 200,000 books into Stanford's collection of more than 7.5 million. Keller shrugs off publishers' concerns that the project is self-serving and short-sighted. He argues that the work will serve as a platform for services that will enhance library users' search capabilities and deepen their engagement with content.

For example, readers will be able to click on books in Stanford's catalogue and be "recommended" other material they might find interesting, based on previous readers' withdrawals. Other tools in the works include taxonomic searching, which allows readers to perform searches for concepts, not just keywords; links in digital documents to footnoted source material; and the ability to click on words to pull up definitions. Keller says that a "pretty worked-out system" comprising this new functionality could be running in five to ten years' time.

Although "the notion of the library as a physical building is only 50 per cent of the idea", Keller says, physical considerations are not overlooked.

Student demands and competition between campuses mean libraries must think imaginatively about how they provide their services.

Keller has overseen extensive renovation at Stanford, instituting multiple "modes of seating" to cater to those who like studying in a "salon setting"

and are undisturbed by a little noise, and to others who prefer absolute silence. Technology is, of course, being woven into the library. TeamSpace, a collaborative software program developed at Stanford, will before long facilitate group work by displaying inputs from multiple laptops on a large screen in a library room.

"The rush of change is so huge," Keller says. "The big challenge for those who might regard librarianship as a plodding profession is to keep up."

É«ºÐÖ±²¥

The conference of the Society of College, National and University Librarians is being held in Newcastle, June 21-23.