Do libraries need more shelving? Isn’t everything digital?


The University of California system, like many research libraries, relies heavily on high-density off-site shelving.  We own and operate two facilities, one each in the north and the south of the state, which combined hold almost half of the combined physical collections of the library systems of all 10 UC campuses.

And both facilities are almost full.  We estimate that we will run out of shelving for regular volumes (we reserve some shelving for unusual sizes and types of materials, and some of that will fill a bit later) in two, perhaps three years.  This has us quite worried: an expansion will cost about $25 million (our design is modular: that’s for a one module expansion, for about 2.5-3 million volumes), and will take some time to design and build: we’re are getting close to the fail point.  Given declining state funding (including a complete cut-off of all capital funding), we are not sure where we’ll get the money or how soon.

A big part of the problem is convincing people that more shelving for print collections is a vital university need in a time of declining budgets.  After all, everything is digital now, right?


One huge misconception we face is that digitizing our collections means we don’t need the print anymore.  For example, we are participants in the Google Books / HathiTrust project, and most of our 11 million regular volumes have been digitized.  Why not burn our print copies?

  1. For starters, about half of the collection is still in copyright.  The HathiTrust collection can be searched, full-text, to find the existence of books, but we are not allowed to let people use the digital copy (with limited exceptions, e.g., for the blind, who can listen to a text-to-voice conversion).  Decades before this need for our print copies goes away.
  2. Second, we are here not to build collections for their own sake, but to serve our faculty and students.  And many of them vastly prefer doing their work from print copies.  Those who read long monographs find it easier and their comprehension higher.  Those who need to study large images or maps, in high resolution, or who want to see side-by-side page comparisons, need the print.  And for many rare and historical documents, the materiality of the original document itself is of enormous importance for scholarship, from the marginal annotations to the construction of the volume.
  3. Next, we can have little or no confidence that we can guarantee long-term digital preservation.  Digital storage has been around a relatively short time.  In that time, formats change frequently.  Hardware and software to render digital formats changes.  Bits on storage media rot.  Keeping bits and being able to find and access them in the future requires large annual expenditures, and those expenditures are getting larger as the amount of content we want to preserve grows enormously fast.  Further, much of scholarly content currently is held on servers of for-profit companies, and we have no guarantee those companies will survive, or that they will take care to ensure that their archives of scholarly publications survive.
  4. The Google project has been very good, but it is not complete.  It does not scan fold-out pages, for example, which are in many scholarly books (maps, charts, tables).  We have discovered that sometimes they miss pages, or the quality is not readable.

So, for now, there is pretty much consensus among research scholars and librarians that we must keep print copies for preservation in all cases, and for continuing use in many cases.

We are able to moderate shelving needs somewhat by reducing the number of print copies that are stored, and we are actively engaged in de-duplication and shared print projects.  But too few copies provides a great risk of irretrievable loss or damage, so this can only (prudently) go so far.

OK, so we can’t eliminate the need for shelving.  But do we need to expand remote shelving?


There are two very simple and compelling reasons. First, many campuses (including several of the UC campuses, like Berkeley, UCLA, San Diego, etc.) were located in areas that have become extraordinarily successful — and thus expensive.  Real estate prices in Berkeley are twice that in Boston for example, and six times higher than in Champaign, IL (I compared residential real estate prices, but the differences are primarily due to underlying land prices, and campuses are located near prime residential real estate in any case). As our campuses need space for new educational and research facilities, they are — quite reasonably — requiring that more of our book shelving go off-site, to cheaper locations: the cost of fast paging and delivery is far less than the real estate savings.  Half of Berkeley’s collection is already stored off-site.  At least one of our campus library spaces is closing in the next couple of years (quite possibly more), and two of our campus libraries were recently transformed into learning spaces without standalone book collections, requiring even more moves to off-site shelving.

The second reason we need more shelving (wherever it is located) is also simple: we need to expand our print collections.  Despite our savings through de-duplication and shared print consortia, an enormous amount of new scholarly research is only published in print, still (this is especially true for foreign language publications, which account for nearly half of our new acquisitions).  And, of course, those students and scholars who demand print copies for their use, demand that of newly published materials too, even if they are available digitally.

I’m a digital guy.  My paper files are almost non-existent.  I almost always read scholarly journal articles on my screen, and for about half the books I read I use my Kindle or MacBook.  But I’m here to witness to a very serious fact facing the 21st century research library: we need more print shelving space, especially off-site shelving.


Can we afford privacy from surveillance? Do we want to?


A couple of weeks before I started my position as University Librarian, the UC Berkeley School of Information invited me to give a talk on the future of individual privacy; here is a video of that talk.  Last week, nationally-syndicated radio show host Katherine Albrecht interviewed me on this topic for about 45 minutes; here is an MP3 of the show (with many commercials, I’m afraid).

In short, I think the economics of surveillance and protection from surveillance are leading inexorably to a not-very-distant future of radical transparency, at least for any information about us that is captured and stored on digital, networked-computers (which is more and more all the time, and will be even more when the Internet of Things really takes off).  I don’t see an alternative: we get to much value from selective revelation of information about ourselves, value that will be increasing as we learn better ways to network and use that information.  And the costs of capturing networked information are going down faster than the costs of protecting ourselves, and I think this is a technologically unavoidable fact driven by the nature of selective revelation in a networked world.

Relevance for libraries?  You might be thinking, “libraries have strong policies to protect the privacy of their users information.”  Yes…sort of.  First, policies are themselves a technology, and they are costly to enforce.  How good is our security against data breaches?  Better than at the IRS, or at JP Morgan Bank?  How fast are our budgets for security growing?

Another issue: to provide our users with access to the rapidly expanding networked stores of information, we provide them with access to an ever increasing array of third-party tools and databases.  What sort of privacy protections do we have on how those third-parties protect our users’ privacy?  Do we have contractual provisions with all of them? No.  (Can you spell “Google”?)  And contractual provisions are another type of policy, that needs ever-increasingly expensive enforcement, whether it be cybersecurity against external attacks, or protection against unscrupulous employees who might sell access.

Do universities need libraries? Isn’t Google free?


On Friday I gave a lunchtime talk to the UC Berkeley Foundation (about 100 people — mostly alumni donors — who manage and lead fundraising on behalf of the campus). I offered an historical framing of just how significant the digital information revolution is going to be (so much more so than we’ve already seen), and why that means we need information professionals more than ever.

I posted the slides to SlideShare, but I tend to use just a few slides to illustrate my points while talking, so I posted them with the script included.

Here is the abstract:

The Gutenberg revolution was an enabler and shaper of the Protestant Reformation, the Renaissance, and the Scientific Revolution. It did so through a small, simple technological advance: merely a reduction in cost and increase in accuracy for information reproduction. But from that modest technological change, one-to-many communication became practical.

The digital revolution accomplished the same feat, only more so: the incremental cost of information reproduction is now about zero; reproduction accuracy is about perfect. And a new impact: information distribution is instant. These are even greater transformations than the Gutenberg press, which enabled and shaped the Protestant Revolution, the Renaissance and the Scientific Revolution. The impacts on civilization of the digital revolution over the coming decades and centuries will be even greater.

Though information is now abundant, finding, evaluating, making sense of and using good information is harder. In the Information Age, we need librarians and other information professionals more than ever.