Perhaps not surprisingly, Mr. Allen and I do a good deal of talking past each other.  He correctly observes that I don’t address the plaintiffs’ motivations for filing suit, and characterizes those motivations in ways that put the plaintiffs in a favorable light.  Not knowing the plaintiffs’ motivations, I have no substantive comment except to note again, as I did when the suit was filed, that one might hope that university-based presses would enlist the academic leadership of their universities in efforts to mediate academic matters before initiating legal action.  To my knowledge, no such efforts were ever made by Oxford, Cambridge, or their distinguished presses.

The central claim in my blog post and PW piece is that the remedy sought by the plaintiffs is inimical to normal practices of teaching and learning in universities, and would put insupportable burdens on the behavior of faculty in their development and delivery of courses.  Mr. Allen’s response to that central substantive claim is to let it slide by with the following commentary:

I won’t attempt in this space to explain how the plaintiff publishers’ proposed injunction, if approved by the court and properly administered by GSU officials, could simplify the task of making fair use determinations and obtaining permissions to use copyrighted material in an effective, timely manner.

In other words, he asserts that my claim is wrong (or worse) but chooses not to give even a hint as to why or how it might be wrong.

Mr. Allen then goes on explain how all of Georgia State’s problems, and presumably everyone else’s, would be solved if libraries would sign up for the Copyright Clearance Center’s blanket license, which is available at a reasonable price.  Maybe so, but because not all publishers are covered by the CCC license, universities would still have to go through the process of establishing the bonafides of all of the works not covered by the CCC license These would include works published by Cambridge University Press, one of the plaintiffs.  Unless the coverage of CCC licenses were universal, faculty would still be subject to the tender mercies of new university bureaucracies that would be charged with acting as copyright police, requiring faculty to provide information that they often would not have.

Furthermore, for most libraries, many of the works covered by the CCC license are also covered by other licenses that that libraries purchase directly or indirectly from publishers.  Thus the CCC blanket approach would result in libraries paying more than once for rights to some works, while still having to do onerous handwork for works that are not covered by the blanket license.  The simple solution is perhaps not so simple.  [It's worth noting that the CCC is helping to underwrite the cost of the lawsuit.]

With all of that, Tom Allen and I agree on two important points.  He states that “when academic copying and distribution of material clearly constitute fair use, permission is unnecessary.”  I couldn’t be more pleased to see such a ringing affirmation of fair use as an integral element of copyright law from the AAP, although I expect that there might be some substantive disagreements over the clear constitution thereof. Mr. Allen also observes that the “ecosystem that binds educators, librarians and publishers … will survive this litigation over copyright infringement.”   I have no doubt that it will, but not on the terms of the plaintiffs’ proposed order in the Georgia State litigation.

Here’s the part that motivates this post:

PW: But Michigan librarian Paul Courant, for example, has argued passionately that Google’s books project offers great public benefits, making millions of long-lost books discoverable and accessible, work libraries could never have done so expeditiously. Doesn’t he have a point?

SV: I’m sympathetic to the expediency argument, but I’m also impatient with it. Courant’s argument is based on two assumptions that I take issue with. First is the assumption that the cost to university libraries would be low. We know now that the cost to libraries has actually been significant, and the benefit has been overstated. We also know now that Google wants to be a bookstore, not a library.

Second, the premise that no one else was ever going to do this is an argument by fiat, a classic fallacy. If we, the people of the world, the librarians of the world, the scholars of the world, the publishers of the world, decide that we should have a universal digital library, then let’s write a plan, change the laws, raise money, and do it right. If we’re going to create this with public resources, let’s do it in the public interest, not corporate interest. There’s nothing wrong with Google pursuing a books project, of course, and, yes, there are benefits. But we have to understand that what Google has created is first and foremost for Google, and I think a lot of people have fooled themselves about that.

I respond to Siva’s two points in order.  First, my support for the project is not based on an assumption that the cost to university libraries would be low, but rather on a calculation that the costs are and have been substantially less than the benefits.  The leadership of two dozen or so major research libraries seem to agree with me.  The unsupported assertion that “we now know that the cost to libraries has been significant, and the benefit has been overstated” is, well, an unsupported assertion.  For the University of Michigan Library, the cost has been material but not overwhelming.  We have used staff time, organizational effort, and there has been some disruption of our activities.  But the benefit has been far larger, in the form of making the content of our collections widely searchable, the public domain content readable by anyone, giving us a backup copy of our collections, and in seeding the HathiTrust, which is a cooperative digital library (not a bookstore) with some fifty-two academic libraries as members and a collection of over eight million volumes that is growing by tens of thousands of volumes a week.  (Check out HathiTrust.org).  Siva is welcome to discount these benefits, but first he should count them.

Second, although it’s an honor to be subject to a “classic fallacy” after all of these years as a college professor, I don’t quite see the fallacy.  Just who, other than Google, has been willing to step up and do the job?  And in what pre-Google fantasyland might we have expected the publishers of the world to show interest in making their backlists  part of a universal digital library?  And why would we believe that we can go to Congress and get improvements in copyright law when every time Congress touches copyright law it gets worse?  (Indeed, I would suggest that the Google books project, by showing the value to millions of citizens of digital access to a large corpus of published work, is more likely to move Congress than the excellent public policy arguments that have been adduced, to deaf ears, by the likes of librarians, Siva, and myself.)

I continue to believe that had Google not embarked on this project, and showed the world that mass digitization of library collections could actually be done, we would still be counting the corpus of digitized work by tens of thousands instead of millions.  To be sure, my assertion here is not subject to proof.  We cannot know what would have happened had Google not gone into the scanning business.  What we do know is that no one else was doing it.  And no one else is doing it.  That’s not a fallacy, but a fact.  Actually, two facts.  Maybe someone else would have done it.  But when?  I’m a lot older than Siva, so I’m the one who gets to be impatient when it comes to providing the riches of the world’s academic libraries to the people of the world.

The University of Michigan Library is a founding member HathiTrust, which brings together the resources and digital collections of a large and growing number of public and private libraries and institutions. The HathiTrust does not claim to be a national library, because it aspires to be much more than a national library. Certainly, HathiTrust’s collections and activities could be part of an effort that addresses the grand challenge Darnton has described.

Darnton suggests that a National Digital Library would make “the cultural patrimony of this country freely available to all of its citizens.”  Exactly what this would mean in practice depends on many things, not the least of which are what we might mean by “cultural patrimony” and “available.”  I suggest that the notion of a national collection based on any nation’s “cultural patrimony” is far too narrow.

The collection of the University of Michigan Library, for example, contains works on papyrus and other ancient media that are millennia older than this country. And down the hall from the Papyrus Collection is the Map Library, which contains printed maps from around the world, as well as the global data that is the foundation of modern geospatial information systems.  It simply doesn’t make sense to divide this country’s cultural patrimony from that of the rest of the world.  (And I have to admit that I’m not wild about the word “patrimony,” either.  Many have suggested that “heritage,” would be better, and I believe that Robert Darnton would accept this as a friendly amendment.)

In any case, libraries don’t distinguish between “our” cultural heritage and the cultural heritage from the rest of the world. Libraries, to the extent that their collection efforts are purposeful—and mostly they are—acquire what is intellectually and culturally important, and what is wanted or needed by their clients. This is reflected in the fact that more than 50 percent of the content in the HathiTrust Digital Library—whose partners as of this writing all reside in the United States—is written in a language other than English.  The same is true of Harvard’s libraries, which suggests that Darnton’s idea of  “cultural patrimony” is more inclusive than the words themselves might seem to indicate.

Even the Library of Congress—a government institution—expresses its mission in broader terms. “The Library’s mission is to make its resources available and useful to the Congress and the American people and to sustain and preserve a universal collection of knowledge and creativity for future generations.” (The italics are mine.)

A universal collection of knowledge. What libraries have always striven to provide, and what library patrons have always desired, has now been made possible by new information technologies.  This is the collection that I believe that Bob Darnton is seeking to build, and just such a collection is being collated by the growing partnership that is HathiTrust.

Darnton says he hopes “that the HathiTrust could somehow evolve to become a fundamental building block of a future digital library,” and he goes on to declare that this would “require the permission of Google.” I’m happy to concur with his ambition and also to report that we already have the permission we need from Google to build such a library. U-M’s original agreement with Google, and Google’s subsequent similar agreements with other libraries, contain clauses that permit us to share our Google-digitized copies with other libraries. HathiTrust rests upon this foundation, and Google’s authorization is explicitly confirmed in Michigan’s 2009 amended agreement with Google.

What Google cannot authorize, because it is not in its purview, is the unfettered circulation of in-copyright digital material. I turn to this, what Darnton calls the “vexed question,” and which he prefers to set aside for a later stage of the National Digital Library initiative, because the solution to this problem, the answer to the vexed question, is the sine qua non of any digital library—local, national, or universal—that aims to make the entirety of its content readily available to all comers.

The clause in the U.S. Constitution that is the foundation for copyright law gives Congress the power “To promote the progress of science and useful arts, by securing for limited times to authors and inventors the exclusive right to their respective writings and discoveries”. Copyright law as it stands is working against the very progress that our founders intended to promote—progress that arguably is itself our national cultural heritage, rather than any collection of material, no matter how comprehensive. Restricted access to digital libraries is only one example of current copyright law’s stifling effect, but it’s an important one. The HathiTrust partners are working on the problem, but it’s going to require a powerful coalition and legislative action to make the requisite changes. I hope the representatives of the foundations, libraries, and cultural institutions that have been gathering to discuss Darnton’s proposal will direct the bulk of their energy and resources into this effort, so we can together give the world the universal digital library it deserves.

Many friends and acquaintances have weighed in on the subject, as one might imagine. The PR apparatus at the University, unsurprisingly, wanted to pitch the removal of the catalog as a symbol of the growth of the digital library, with all of the forward-looking and progressive connotations that such a view might support. Most people, librarians and others, were surprised that we still had the card catalog. This included senior faculty who often opine as to how much they love browsing in the stacks but who haven’t actually been there for a decade or two. Younger people didn’t even know what it was. (The current graduating class, it is worth noting, were mostly not born when we retired the catalog.) Most of the library staff were happy to see the thing go. My own view turns out to be at the sentimental end. Forty-some years ago, card catalogs gave me a window on the world of scholarship that left me (and still leaves me) in awe. Of course we have the same information on line, and one is still awed by the resources of libraries great and small. But just as I will always remember Ebbets Field as the location of the first baseball game that I saw in person, the Queen Mary as my introduction to transatlantic travel, and the 6th edition of Samuleson’s introduction to Economics as, well, my introduction to economics, I’ll always remember the card catalog as the rich, powerful and brilliant piece of scholarship that it was, and as a place that I visited in eager anticipation of learning something new. I don’t think that I was ever disappointed.

And then came the plagues.

The week before the catalog was moved, I went down to the basement of the library to find my own cards and those of members of my family, with the intention of removing them and keeping them, why I don’t know. And I’ll never find out. I worked my way through the alphabet, to a box whose last drawer ended with “Cooper.” I figured that “Courant” couldn’t be very far away, but the next case started with “Da-” Where were the Courant cards? It turned out that some years ago there had been a water leak, undiscovered for weeks, that had led to water damage and subsequent mildew and mold. The cards had been destroyed at that time.

On March 8, the cases were trucked to Property Disposition. About half way through the exercise a case hit a sprinkler head, creating a torrent of rusty water, water that looked like blood and smelled worse, damaging hundreds of books (all of which were saved by our overwhelmingly competent preservation staff) and making quite a mess.

With Passover coming, I’m on the lookout for frogs.

Over the past couple of years, The University of Michigan Library has been working on a project to improve the accessibility of our digitized texts for visually impaired UM students, staff, and faculty. First, the team made accessibility improvements to the standard public interface for the HathiTrust Digital Library (formerly known as MBooks) and developed a text-only interface geared toward people with print disabilities that is optimized for screen reading software. Next, and most important, the Library figured out how to grant access to the full text of digitized books for qualified patrons, regardless of the book’s copyright status.

Like many other universities, the UM Services for Students with Disabilities (SSwD) has long offered book digitization service to students with disabilities upon request. This is explicitly allowed under section 121 of U.S. Copyright law.

Our new system basically does the same thing but on a much larger scale. The HathiTrust Digital Library currently provides access to over 4 million digitized volumes and will grow to over 10 million – visually impaired students will have full-text access to all of these volumes. We consider this just the beginning. Over the next year, we will continue to work on improvements to the interface and conduct more user assessments, and our HathiTrust partners are working together to create a framework through which we can offer this service to users at their institutions.

Once a University of Michigan student registers with the UM Services for Students with Disabilities any time she checks out a book that has been digitized, she will automatically receive an email with a URL. Once the student selects the link, she will be asked to login. The system will check to see whether the student is registered with SSwD as part of this program, and ensure that she has checked out this particular book. If the student passes both of those tests, she will get access to the entire full-text of the book, whether it is in copyright or not, in an interface that is optimized for use with screen readers. The Library’s Blog for Library Technology has more details on the technical elements for those who are interested.

Our system was endorsed by the National Federation for the Blind as a model for how libraries can serve visually impaired patrons in the digital age. It’s a great example of how digital technologies can extend the ability of libraries to serve their clients, and to extend learning and teaching beyond traditional populations.

As with the production of so many things that are of broad public value, this work could not have happened without the efforts of a committed champion. I am pleased to recognize the commitment and skill of Jack Bernard, an attorney in Michigan’s Office of the General Counsel, who has provided substantive and legal leadership in our efforts to make our collections accessible to our students with print disabilities. Jack’s efforts were recognized by the American Library Association, which gave him the L. Ray Patterson Copyright Award in 2009.

The Economist interviewed me about the settlement at some length, and made a podcast that I quite like. It recapitulates fairly painlessly (it’s 13 minutes) some of things that I’ve been saying about the Google lawsuit and settlement for some time.

And, for something completely different and arguably more important, Paul Krugman has a superb piece entitled “How Did Economists Get It So Wrong” in the New York Times Magazine of September 6. What’s remarkable is how economists got it so wrong 70 years after Keynes got it so right. Anyhow, this is a testimonial for Krugman’s piece from an admiring economist.

I learned something important that at some level I already knew.

The most important single thing about the Google settlement, simultaneously its greatest achievement and among its most vexing features, is the treatment of orphaned works (in James Grimmelman’s witticism, “zombie” works). The problem, as we all know, is that there are millions – no one quite knows how many – of works that may or may not be in copyright and for which the rightsholder(s) may or may not exist and may or may not be aware of their rights. Our ability to use these works is thus much compromised: we run the risk that a copyright holder will appear and claim damages. As we all know, Congress’s efforts to make it easier and safer to use orphaned works have failed. Moreover, the most recent draft legislation would have imposed difficult and costly burdens on a potential user by requiring the would-be user to make substantial efforts to find any potential but unknown rightsholder.

Along comes the Google settlement, which solves at least part of the problem, for Google and the Book Rights Registry, at one fell swoop. (Only part of the problem, because works that were not registered with the copyright office will likely not be in the settlement and yet may be just as orphaned as those that are registered.) Under the settlement, revenues generated by orphaned works will be held in escrow for for five years, allowing time for a rightsholder to come forward. It’s a moving window; if the rightsholder comes forward in year 22, she gets revenues from year 17 on. Thus the products that Google sells to individuals and institutions can include, among other works, millions of orphans (zombies). Without the orphans, the great public benefit of the settlement – the ability to find and use much of the literature of the 20th century in digital form – would be much diminished.

At the same time, the disposition of the revenues attributed to orphaned works is one of my least favorite parts of the settlement. The unclaimed revenues go first to support the operations of the BRR, and then, after that, will be used for charitable purposes consistent with the interests of publishers and authors. As the head of a library that has lovingly cared for these works for decades, the notion that the fruits of our labors (and those of many others in many libraries) redound to the benefit of entities that did not write, publish, or curate these works sticks a bit in my craw. So I hope that authors, publishers, the court, and the public will be vigilant in making sure the BRR does not squander the unclaimed revenues on mismanagement, high salaries, and the like. The “charitable purposes” should be an objective, not a remainder for unclaimed funds.

The settlement also gives Google and the BRR, and no one else, the right to use the orphaned works in this way. A number of commentators, have noted problems that may arise from Google’s privileged position in this regard. But there is an obvious solution, one that was endorsed at the Columbia meeting by counsel for the Authors Guild, the AAP, and Google: Congress could pass a law, giving access to the same sort of scheme that Google and the BRR have under the Google Settlement to anyone. And they could pass some other law that makes it possible for people to responsibly use orphaned works, while preserving interests for the missing “parents” should they materialize. Jack Bernard and Susan Kornfield have proposed just such an architecture to “foster” these orphans. Google has also made a proposal that would be a huge improvement.

Given that the parties to the suit, libraries, and the public would all benefit from such legislation, it should be a societal imperative to pass it. I look forward to AAP, the Authors Guild, and Google lobbying and testifying in favor of such legislation. I’d be happy to be there, too.

—–

To the editors:

My colleague and friend Robert Darnton is a marvelous historian and an elegant writer. His utopian vision of a digital infrastructure for a new Republic of Letters (Google and the Future of Books, NYRB Feb. 12) makes the spirit soar. But his idea that there was any possibility that Congress and the Library of Congress might have implemented that vision in the 1990s is a utopian fantasy. At the same time, his view of the world that will likely emerge as a result of Google’s scanning of copyrighted works is a dystopian fantasy.

The Congress that Darnton imagines providing both money and changes in law that would have made out-of-print but in-copyright works (the great majority of print works published in the 20th century) digitally available on reasonable terms showed no interest in doing anything of the kind. Rather, it passed the Digital Millennium Copyright Act and the Sonny Bono Copyright Term Extension Act. (More recently, Congress passed the Higher Education Opportunity Act, which compels academic institutions to police the electronic environment for copyright infringement). This record is unsurprising; the committees that write copyright law are dominated by representatives who are beholden to Hollywood and other rights holders. Their idea of the Republic of Letters is one in which everyone who ever reads, listens, or views pretty much anything should pay to do so, every time.

The Supreme Court, which was given the opportunity to limit the extension of the term of copyright, which was already far too long (like Darnton, I think that 14 years renewable once is more than enough to achieve the purposes of copyright) refused to do so (with only two dissenters) in Eldred v. Ashcroft, decided in 2003. Instead, it upheld legislation that, contrary to the fundamental principles of copyright, provided rewards to authors who are long dead, preventing our cultural heritage from rising into the public domain,

In short, over the last decade and more, public policy has been consistently worse than useless in helping to make most of the works of the 20th century searchable and usable in digital form. This is the alternative against which we should evaluate Google Book Search and Google’s settlement with publishers and authors.

First, we should remember that until Google announced in 2004 that it was going to digitize the collections of a number of the world’s largest academic libraries, absolutely no one had a plan for mass digitization at the requisite scale. Well-endowed libraries, including Harvard and the University of Michigan, were embarked on digitization efforts at rates of less than ten thousand volumes per year. Google completely shifted the discussion to tens of thousands of volumes per week, with the result that overnight the impossible goal of digitizing (almost) everything became possible. We tend to think now that mass digitization is easy. Less than five years ago we thought it was impossibly expensive.

The heart of Darnton’s dystopian fantasy about the Google settlement follows directly from his view that “Google will enjoy what can only be called a monopoly … of access to information.” But Google doesn’t have anything like a monopoly over access to information in general, nor to the information in the books that are subject to the terms of the settlement. For a start (and of stunning public benefit in itself) up to 20% of the content of the books will be openly readable by anybody with an Internet connection, and all of the content will be indexed and searchable. Moreover, Google is required to provide the familiar “find it in a library” link for all books offered in the commercial product. That is, if after reading 20 percent of a book a user wants more and finds the price of on-line access to be too high, the reader will be shown a list of libraries that have the book, and can go to one of those libraries or employ inter-library loan. This greatly weakens the market power of Google’s product. Indeed, it is much better than the current state affairs, in which users of Google Book Search can read only snippets, not 20% of a book, when deciding whether what they’ve found is what they seek.

Darnton is also concerned that Google will employ the rapacious pricing strategies used by many publishers of current scientific literature, to the great cost of academic libraries, their universities, and, at least as important, potential users who are simply without access. But the market characteristics of current articles in science and technology are fundamentally different from those of the vast corpus of out-of-print literature that is held in university libraries and that will constitute the bulk of the works that Google will sell for the rights holders under the settlement agreement. The production of current scholarship in the sciences requires reliable and immediate access to the current literature. One cannot publish, nor get grants, without such access. The publishers know it, and they price accordingly. In particular the prices of individual articles are very high, supporting the outrageously expensive site licenses that are paid by universities. In contrast, because there are many ways of getting access to most of the books that Google will sell under the settlement, the consumer price will almost surely be fairly low, which will in turn lead to low prices for the site licenses. Again, “find it in a library,” coupled with extensive free preview, could not be more different than the business practices employed by many publishers of scientific, technical and medical journals.

There is another reason to believe that prices will not be “unfair”, which is that Google is far more interested in getting people to “google” pretty much everything than it is in making money through direct sales. The way to get people to come to the literature through Google is make it easy and rewarding to do so. For works in the public domain, Google already provides free access and will continue to do so. For works in the settlement, a well-designed interface, 20 percent preview, and reasonable prices are all likely to be part of the package. Additionally, libraries that don’t subscribe to the product will have a free public terminal accessible to their users. This increases the public good deriving from settlement both directly and by providing yet another distribution channel that does not require payment to Google or the rightsholders.

The settlement is far from perfect. The American practice of making public policy by private lawsuit is very far from perfect. But in the absence of the settlement – even if Google had prevailed against the suits by the publishers and authors – we would not have the digitized infrastructure to support the 21st century Republic of Letters. We would have indexes and snippets and no way to read any substantial amount of any of the millions of works at stake on line. The settlement gives us free preview of an enormous amount of content, and the promise of easy access to the rest, thereby greatly advancing the public good.

Of course I would prefer the universal library, but I am pretty happy about the universal bookstore. After all, bookstores are fine places to read books, and then to decide whether to buy them or go to the library to read some more.

Paul N. Courant

Note: This letter represents my personal views and not those of the University of Michigan, nor any of its libraries or departments.

The books will come from a number of academic libraries, including the University of Michigan, the University of California, and Stanford University, which have been participants Google Book Search from the beginning, These three worked with Google during the settlement negotiations in an effort to shape the settlement to serve the interests of research libraries and the public, as discussed in a joint press release.

The settlement is complicated, and as people work through it I expect a lively set of discussions and I invite comment on this blog and elsewhere. I’d like to start with what I see as a couple of key points.

First, and foremost, the settlement continues to allow the libraries to retain control of digital copies of works that Google has scanned in connection with the digitization projects. We continue to be responsible for our own collections. Moreover, we will be able to make research uses of our own collections. The huge investments that universities have made in their libraries over a century and more will continue to benefit those universities and the academy more broadly.

Second, the settlement provides a mechanism that will make these collections widely available. Many, including me, would have been delighted if the outcome of the lawsuit had been a ringing affirmation of the fair use rights that Google had asserted as a defense. (My inexpert opinion is that Google’s position would and should have prevailed.) But even a win for Google would have left the libraries unable to have full use of their digitized collections of in-copyright materials on behalf of their own campuses or the broader public. We would have been able, perhaps, to show snippets, as Google has being doing, but it would have been a plain violation of copyright law to allow our users full access to the digitized texts. Making the digitized collections broadly usable would have required negotiations with rightsholders, in some cases book by book, and publisher by publisher. I’m confident that we would have gotten there in time, serving the interests of all parties. But “in time” would surely have been many years, and the clock would have started only at the end of a lawsuit that had many years left to run. Moreover, each library would have had to negotiate use rights to its own collection, still leaving us a long way from a collection of digitized collections that we could all share.

The settlement cuts through this morass. As the product develops, academic libraries will be able to license not only their own digitized works but everyone else’s. Michigan’s faculty and students will be able to read Stanford and California’s digitized books, as well as Michigan’s own. I never doubted that we were going to have to pay rightsholders in order to have reading access to digitized copies of works that are in-copyright. Under the settlement, academic libraries will pay, but will do so without having to bear large and repeated transaction costs. (Of course, saving on transaction costs won’t be of much value if the basic price is too high, but I expect that the prices will be reasonable, both because there is helpful language in the settlement and because of my reading of the relevant markets.)

The settlement is not perfect, of course. It is reminiscent, however, of the original promise of the Google Book project: what once looked impossible or impossibly distant now looks possible in a relatively short period of time. Faculty, students, and other readers will be able to browse the collections of the world ‘s great libraries from their desks and from their breakfast tables. That’s pretty cool.

The choices that we face are indeed less than perfect. We can choose purity and perfection, and not permit any restrictions on the use of scans of public domain material, with the result that the rate of scanning and consequent display will be pitifully slow. Or we can permit corporate entities, including dreaded Google, to scan our works, enabling millions of public domain works to be made available to readers on line, at no cost to the readers, in a relatively short period of time. I am on record by word and deed as preferring the second choice.

In his paper, Johnson notes that the original works are retained by the libraries and could be scanned again. He fails to note that libraries whose PD works are scanned by Google get to keep a copy of the scans and are free to display them on line, independent of Google Book Search. Over 300,000 public domain works can be found in the University of Michigan catalog and read on line. The number grows by thousands per week. Of course I would prefer it if the digital files could be used without restriction. Would someone please tell me the name of the entity that stands ready to digitize our collections, for free, without restriction on the use of the digital files? In the meantime, it seems to me that making the books available to readers online makes for a better world, albeit, sadly, not a perfect one.

And, this just in, an article by Kalev Leetaru in First Monday that compares Google Book Search and the Open Content Alliance and finds much that is both good and less than perfect in both.