PaperPersists: Why Physical Library Collections Still Matter

by Walt Crawford

ONLINE, January 1998
Copyright © Online Inc.

There are enormous economic and ecological disadvantages to the all-digital library.
Last week, Patrick Hogan called from ALA Editions. "The project's been approved, and the written contract is in the mail." The contract is for a new book on the future of media and libraries that I'll start writing as soon as I finish this article; the book should appear in time for the American Library Association's 1999 Midwinter Meeting. Why would ONLINE readers care about this? I'd like to think that some of you will buy the book, but more interesting, given projections of the past years, is the mere fact of this transaction leading to a printed book.

I'm a lifelong technologist who's been on the Internet since the late 1980s. I make my living designing and promulgating services that run on the World Wide Web. I should know better than most that print is dead, the book is obsolete, the future belongs entirely to digital transmission, and the screen's the place for reading. This is not going to be a quickie book about using the Web, writing Java, or any of those hot topics. So why am I contributing to a dead medium?

Because paper persists. A paperback book is the best way for me to communicate a fairly lengthy and complex narrative discussion.

Paper persists. The physical print collections in public and academic libraries will continue to grow and be central to the missions of those libraries.

Books continue to matter, now and for any plausible future. Not as the only means to transmit information, entertainment, and knowledge--that hasn't been true for more than a century. Not as the dominant force among media--that hasn't been true for decades. But as a vibrant, healthy medium--one that serves a variety of needs better than any alternative and that makes good economic, ecological, and technological sense for the new millennium--the book just isn't going away.

Neither are print magazines, such as the one you're reading now. Magazines and journals are very different sets of media with different strengths, characteristics, and problems; the very real problems of scholarly journals don't have much to do with the future of magazines--but that's another topic, to be taken up some other time.

If your first reaction to this article's title was "Well, of course" then you're in the mainstream for 1998--but you might have been considered a Luddite among high-ranking librarians of the 1980s and early 1990s. If your reaction is "That's stupid. Print is dead: that's inevitable" then you'll dismiss this article anyway, and might as well turn the page now. Outside the library profession, prophecies of the death of print and the all-digital future began in the late 1980s, reached a peak around 1992-1994, and are now declining into the oblivion they deserve. Within the field, however, the "common knowledge" that print was dying seems to go back much further, typically as part of a clarion call for libraries to reinvent themselves for the all-digital future.

A PowerPoint Fable

What follows is a lie. There was no single meeting at which all the great library minds concluded that the future should be all digital, and PowerPoint didn't even exist in 1977. But let's pretend there was a grand summit meeting, say as a preconference to the American Library Association's 100th Anniversary annual meeting that year. All the top people in libraries and library schools attended. For the keynote at this meeting, a great guru of libraries and technology presented the following as a slowly building and absolutely compelling PowerPoint presentation. You can imagine the bullets sweeping onto the screen, accompanied by suitable music and fireworks. Here's the final screen:

  • When reading from digital devices is as comfortable, effective, and fast as reading from printed books and serials (or more so), and

  • When digital reading and storage devices are omnipresent, and

  • When digital distribution replaces print publishing for all new materials because it's cheaper, faster, and better, and

  • When all existing library materials are converted to digital form, and

  • When digital communications facilities are so fast and inexpensive that transmission of publication equivalents is essentially instantaneous and free, and

  • Given that publishers won't stand in the way of institutions making single purchased or converted publications simultaneously available throughout the institution, nation, or world, then:

  • Libraries will and must convert to digital distribution as a more effective way to carry out their missions.

That all makes pretty good sense--although I would argue that most of those "whens" should be "ifs." In the early days, some of them probably were stated as "if and when."

But then two things happened, over a period of years:

  1. The "ifs" became "whens."

  2. People remembered the conclusion without remembering all the premises.
These assumptions grew during the 1970s and early 1980s, remarkable times for academic libraries and technology. Those were growth years, when everything seemed possible and everything desirable seemed almost inevitable. Barriers of money and technology were scarcely barriers at all. Technology lowered the money barriers, and grant or government funding took care of the rest.


One trend that began in the eighties has continued far beyond its useful life. That is the great technological handwave--the futurist's response to any shortcomings in technology, any unmet needs, anything that's lacking. When you hear, for example, "we can confidently project that such devices will be commonplace in the next two years," you're hearing the great technological handwave. Flat screens with better resolution and readability than printed pages? Two years from whenever you ask. High-speed interactive communication for every household at costs so low nobody will notice? Well before the end of the century. And so on.

The great technological handwave turns "ifs" into "whens" and "whens" into "just a couple more years." The great technological handwave rejects budgetary arguments, since as we all know technology just keeps getting cheaper and cheaper until it's essentially free.

The cousin of the great technological handwave is that magic word "inevitable." The great digital convergence? Inevitable. The death of print? Inevitable. Which is another way of saying, "you probably won't like this, and I can't make a compelling case." If the case is strong, the I-word is pointless. These days, inevitability is invoked whenever one questions the common assumptions of the past. What once was desirable is now inevitable. When you hear "inevitable," substitute "rowrbazzle." It means as much and it's more fun to say.

As for the great technological handwave? Technological improvement is neither smooth nor entirely predictable, even vastly-improved tools rarely catch hold immediately, and most new devices and techniques never become important. Paraphrasing Paul Saffo, it's true enough that we tend to overestimate the short-term significance of new technologies and underestimate their long-term significance--but Saffo fails to point out that, perhaps 80 percent of the time, the new technologies simply disappear or fade into specialized use.

Unless you really did fly to work in your personal helicopter from your solar-powered household, or sat back as your car drove itself on today's digitally-controlled freeways, you would do well to treat the great technological handwave with a smile and deep distrust.


What happens if the premises arguing for library conversion to digital fail? Logically, if the premises are invalid, then the conclusion is false or at least unsupported. So we must either validate each premise or conclude that the premise really isn't necessary; otherwise, the argument for converting to digital libraries is without foundation. Let's examine each "when" from that mythical 1977 PowerPoint presentation.

Reading from Digital Devices

If every long text is printed out each time it is used... a typical public library would spend much more on printing and licenses than its current total budget and would use at least 50 times as much paper as at present.
One absolute article of faith in the 1970s, 1980s, and early 1990s was that the DynaBook, or its equivalent, was just around the corner. This device offers better readability than a book and easier navigation. It is light enough in weight and has a high enough battery life so that it is as portable as a book; with rapid replacement of contents, it functions as a universal book. Every projection I've seen had such a device on the market long before now, at an extremely modest price.

It hasn't happened, and there's every reason to believe that it won't. Reading from digital devices, whether portable or desktop, suffers in several areas--among them light, resolution, speed, and impact on the reader--and there has been essentially no improvement in any of these areas in the last five years.

Many futurists have conceded this point. They now admit that people will print out anything longer than 500 words or so. It's just too hard to read from a computer, and it doesn't seem likely to get a lot easier. If every long text is printed out each time it is used, there are enormous economic and ecological disadvantages to the all-digital library: briefly, a typical public library would spend much more on printing and licenses than its current total budget and would use at least 50 times as much paper as at present.

Omnipresent Electronics

The public as a whole has no need for--or interest in--digital book equivalents.
What ever happened to Sony's BookMan, their portable digital book? Why didn't the DynaBook ever emerge as a real device? Why aren't we all using Personal Digital Assistants for most of our reading? The answers are complex, but the overall situation is clear. The PDAs being produced today and designed for tomorrow aren't intended to function as book replacements: the screens are small, hard to read, and awkward to navigate for lengthy text. It's increasingly clear that the public as a whole has no need for--or interest in-- digital book equivalents.

Two-thirds of adult Americans, and a higher percentage of children, use their public libraries. Roughly two-thirds of adult Americans purchased books last year. I'd guess that an even higher percentage reads magazines or newspapers. Is it possible that electronic tablets could achieve such ubiquity in the next few years--or even the next couple of decades? I doubt it.

The Death of Print

We've heard about the death of print for years now, too often from within the library field. While print has been dying, the publishing industry has been growing. As a long-term trend, more books are being published and purchased, more issues of magazines are being circulated, and more revenue is making a substantial industry even larger.

One academic library expert anticipated in 1992 that the market for information printed on paper would shrink by 50 percent within five years. By the end of 1997, the market for information printed on paper was substantially larger than in 1992.

Print publishing is actually several related industries, most of which are healthy and growing. For a variety of reasons, not the least being people's preferences, I don't see the situation changing soon. Publishers don't spend much time these days talking about the death of print--that was last decade's news. They certainly hope to be part of the new markets that complement print, but they know print isn't going away.

Universal Conversion

The Library of Congress continues to acquire new print materials much faster than it digitizes old ones.
When will all existing library materials be converted to digital form? Not in my lifetime, probably not in yours, and quite likely never. The task is too big and too expensive, and the reward keeps diminishing.

The book I wrote with Michael Gorman, Future Libraries: Dreams, Madness & Reality (ALA Editions, 1995), deals with conversion questions in some detail. The digital conversion efforts being mounted by the Library of Congress and a cluster of university libraries provide strong indications of what's happening. To wit, collections of material will be digitized, primarily material that can't be made available otherwise: unique photographs, manuscripts, brittle books of unique importance, and the like. My organization, the Research Libraries Group, is involved in such efforts, as are many others. These projects will yield digital collections that enhance and extend libraries. They will not yield all-digital libraries, and there's no indication that such efforts would ever scale up to complete conversion.

The Library of Congress continues to acquire new print materials much faster than it digitizes old ones. If anyone still has universal conversion as a goal--which I doubt--we're moving backwards.

Digital Communications

We come now to the fifth "when," and it's a doozy: when communications are essentially free... Who's providing those unlimited pipelines? Who's paying for the technical support to keep them operational? Where did we ever come up with such nonsense as "essentially free"?

"Essentially free" is another way of saying "phenomenally expensive, but the incremental cost becomes small." "Essentially free" is a technological handwave. It's always wrong. If it is possible to build an international network that could actually provide everyone with universal multi-way video-speed communications capabilities, from any point to any point, it would probably cost hundreds of billions of dollars--and as projections of the possible revenue become more realistic, the will to spend that money vanishes.

"Essentially free" is essentially nonsense. Yes, a $2500 PC purchased today is some 75 times as powerful as the $2500 PC of 1988-- but that doesn't mean you can buy a useful PC for $33! Technology doesn't work that way; increased performance for a price doesn't mean that prices keep going down for acceptable performance.

Publishers and Rights

It's unreasonable to expect publishers to survive--or the editorial, acquisition, publicity, and other publishing functions to continue--if they face the possibility of a single copy being sold, then distributed universally.
Finally, there's that crucial "given:" that digital resources won't raise new issues of rights, payment, and so on. Once a library buys a book or sound recording, they can lend it out over and over, with no further payment required. It's supposed to be better yet with digitized materials--many people can read or use the item simultaneously. What a cost savings!

But it won't happen that way. The Association of American Publishers, for one, has made it clear that their view of digital resources in libraries is strictly pay per view, with libraries essentially serving as distributors for the publishers. The publishers' attitude is justified. It's unreasonable to expect publishers to survive--or the editorial, acquisition, publicity, and other publishing functions to continue--if they face the possibility of a single copy being sold, then distributed universally. For that matter, what non-academic author will write if the total compensation is royalty on a single copy?

It's not a sufficient answer to say that digital resources can be made available with simultaneous user restrictions, so that only one or two readers can use them at a time. For online indexes, that methodology makes perfect sense; it's how RLG typically sells Eureka and Zephyr services, for example, and increasingly how other online resources are sold.

But what do simultaneous-use restrictions mean for digital book replacements, where users will read from printed copies? The restriction lasts long enough for a reader to download the file; then another user can happily download it while the first is printing it locally. The net effect is still publication without repeat sales.

What the publishers want is a hefty fee each time a file is touched or downloaded or examined to any real extent. Is that unreasonable? Perhaps not, where end-users are concerned. When you add the cost of printing to the download fee, the end-user's cost is likely to be comparable to what they'd pay to buy the item now (although the resulting stack of paper won't be as convenient as a book). But for libraries, as compared to circulating or reference collections of published books, it's a terrible idea.

Publishers have accepted (if not always graciously) the idea that one copy of a book will be read sequentially by quite a few different people. They produce mass-market paperbacks to encourage people to buy copies of some books, at a price less than that of photocopying a library book and yielding a much more convenient (and legal) product. This continuing balance has worked to the mutual benefit of publishers and libraries.

In an all-digital age, things would be different--and I see no likelihood that the differences would favor libraries. Publishers issue papers calling for cooperation with libraries, but only on publishers' terms. And, I say again, those terms are at least partly reasonable.

Six Strikes, No Hits

That's the last of the premises leading up to the inevitability of all-digital libraries. Do such libraries still seem inevitable, in the foreseeable future? Not to me--at least not for public and academic libraries. There will certainly be some special libraries that lack central physical collections and still serve their users well--indeed, such "virtual libraries" already exist.


You still see projections of an all-digital future and the coming irrelevance of print collections, but those projections become ever more other-worldly as time goes on. Within the library field, such assertions seem to come primarily from three sources: Outside librarianship, most such projections come from futurists and self-serving prophets, who rely on the comforting fact that nobody seems to care how often a futurist is wrong, as long as he or she is interesting.

Why did so many futurists and technologists proclaim the death of print in the past, and why do some still do so? A variety of motives present themselves, including the simplification that comes with projecting visions for the future; taking narrow perspectives on technological possibilities, without considering broader issues; assuming that everyone has the same fascination with the new and impatience with the old that some futurists exhibit; and even the need to be controversial.

Whatever the reasons, some still proclaim that the all-digital future is just around the corner, ready to sweep away all traditional media in a grand, enlightening convergence. Some will even tell us that history is dead--that the history of new media altering but rarely obliterating older media, the history of people adding new choices without eliminating older choices, the history that shows most technological innovations simply failing, that all this history is simply irrelevant.

Since claims that history is dead cut off logical discussion just as surely as claims of inevitability, the only reasonable response to such claims is: "Nonsense. People are still people, and neither history nor natural laws have been repealed."

Sensible thinkers have pretty much abandoned the death-of-print school. When Bill and Melinda Gates established the Gates Library Foundation last year, they deliberately used the word "library" and asserted that books would continue to be at the core of good libraries in the future, even as they help libraries to expand their resources beyond books. Indeed, some editors of Wired magazine, home to one of the most assertive death-of-print claimants, have admitted that when complex thoughts need to be understood, boring old books work best.


One tenet of the all-digital believers is that all that counts is information, and that information is more malleable and available when it's handled digitally. Further, the volume of information in the world doubles every five years, and the amount of digital information is growing so rapidly that print occupies a narrow niche in comparison. These are both at best half-truths. "Information" has become a bland, almost meaningless word. In many cases, what's really meant is data: that is, raw facts. That's certainly true in the case of "doubling every five years."

No doubt the amount of raw data generated and stored does double every five years. It's certainly true that the number of characters printed in books in any given year is trivial compared to the number of bytes of raw data generated in that same year. That's been true for some time now; it's also entirely irrelevant. It not only compares apples with oranges, it compares pineapples with pole-vaulters.

A simple calculation will show how trivial the volume-of-raw-data argument really is. Let's say that 1.5 billion books are printed in the United States in a given year. At an average of 600,000 characters per book, that's 900 terabytes of "data." Now consider the broadcast of a three-hour World Series game seen by ten million people. Given that a standard TV signal uses a bandwidth of 5MHz, equivalent to 625,000 characters per second, those viewing that broadcast receive, in total, 6,750 terabytes: seven times as much as all the books for that year, in a single World Series game! (Similarly, more "information" enters my house each day on the 80-channel TV cable than I will ever read in books and magazines throughout my life.)

To which the only reasonable response is: "So what?" Raw data isn't information, and the fact that each television viewer sees "one book's worth" of raw data each second is irrelevant for any rational purposes.

Information requires mental or mechanical processing--it requires organization and context to make it at least minimally meaningful.

For that matter, books aren't entirely about information, any more than public libraries are entirely about information. Books are about meaning, knowledge, understanding, wisdom, and narrative: words given life. Quoting from Future Libraries: Dreams, Madness & Reality:

Knowledge can be defined as information transformed into meaning... Understanding is knowledge integrated with a world view and a personal perspective and exists entirely within the human mind, as does wisdom, understanding made whole and generative.
What of narrative? I use narrative as shorthand for literature of all sorts, including prose, poetry, nonfiction and fiction. There seems little doubt that narrative communicates best in printed form--that the interaction between author and reader, with the reader bringing his own mental images into play, works best in print.


I've heard plaintive cries along these lines: "But these technologies are so neat, and they would offer so much, why doesn't their inevitable triumph happen faster?" Perhaps because people aren't machines, and don't respond mechanically. The advantages of the all-digital future are primarily mechanical advantages; the disadvantages prevail when people enter the equation. That's not an easy problem to solve, since eliminating the people tends to eliminate the market as well.

True digital zealots will accuse me and other "book lovers" of being irrational--but a preference for a medium that works beautifully over one that works less well is not only rational, but eminently sensible.

It's been demonstrated repeatedly that people will pay to have good public libraries: most bond and tax override elections succeed. Those who have talked to the people consistently get a more refined message: what people want from their libraries most, and what they'll most readily pay for, is a good supply of books. Almost any public librarian in touch with their community, no matter how rich, advanced, or sophisticated the community, will have one answer to the question: "What single thing are your customers most ready to pay for?" The answer is always books--not "information" but books. Librarians ignore that answer at great peril.

Paper Persisting

Some areas of book publishing have already been replaced (partly or wholly) by digital publications and distribution. That's as it should be. I shed no tears when Readers' Guide to Periodical Literature is replaced by online and CD-ROM indexes. Most households are better served by $30-$50 CD-ROM encyclopedias updated every year or two than by massive $600-$1500 print encyclopedias that the households will never update. While good print atlases continue to matter, the best CD-ROM atlases offer ways of seeing the world that a print atlas can't duplicate--and these days, most newer CD-ROMs also point to newer updates on the World Wide Web.

One thoughtful analyst suggests that half of what's now printed in book form (that is, sheets of paper somehow bound into volumes) could be replaced by digital publishing and distribution, even as bookstores and public libraries thrive. How? By replacing most or all of the parts catalogs, operating manuals, maintenance guides and similar items that aren't meant to be read end to end, but require brief consultation from time to time. Such replacement should help traditional book publishing by reducing the demand for paper, ink, and press time, and by further weakening the already-questionable ecological arguments against printed books. These changes would further increase the ratio of stuff that's only in digital form to stuff that's in printed form, but that's a pointless ratio in any case.

Books matter, and will continue to matter, because people learn from them and enjoy reading them. Public and academic libraries will continue to rely heavily on printed collections because they work so well for the ideas of the future as well as the record of the past and present. Of course, libraries will extend those printed collections with in-house media collections, borrowed physical resources, and an ever-growing array of digital publications and online retrieval: that's neither revolutionary nor even new.

The "And, Not Or" Vision

Visions are wonderful things, though frequently misleading. I've always disclaimed any thought of being a visionary--but, as it turns out, I do have a vision of sorts. That vision is embodied in the phrase "And, Not Or," which I first used in fall 1992, when I began speaking out about the future of libraries and print. That vision can be summarized as follows, excerpted from a longer version in Future Libraries: Dreams, Madness & Reality: That's my vision for the future. It's a vision that will enter into future articles about aspects of media and libraries. And it's a vision that calls for strong printed collections at the heart of every public and academic library, just as strong libraries function at the heart of every community, college, and university.

Author's Note: Portions of this article (specifically the PowerPoint fable) first appeared in different form in speeches delivered in late 1996 and early 1997. If you were expecting a personal computing article here, don't worry--I'll still be writing those, mixed roughly half and half with essays on the future of media and libraries. Although you can't place advance orders for my next book yet (its current working title is Being Analog: Building Tomorrow's Libraries), Future Libraries: Dreams, Madness & Reality continues to be great reading--and it won't be replaced by the new book.

--Walt Crawford

Communications to the author should be addressed to Walt Crawford, Research Libraries Group, 1200 Villa Street, Mountain View, CA 94041-1100; 650/691-2227; Fax 650/964-0943;

[Online Inc.] [ONLINE Home] [Current Issue] [To Subscribe] [Top]

Copyright © 1998, Online Inc. All rights reserved.
[This site created for best results under Netscape.]