Last week, I gave a talk at the California Academic Research Librarians IT conference, "A Workshop on Next Generation Libraries." My Berkeley colleague, Ray Larson, also contributed, and gave an excellent talk that I hope to be able to blog about soon.

I gave a reprise of a talk that I had given when I had just barely started at the DLF; I don't normally repeat talks, but this was a by-request, and so I only modestly modified it. I was explicitly told to "please keep in the parts that are critical of libraries" and so I did, although my views are more nuanced than they were six months ago.

Inevitably the question came up (actually, in multiple guises and iterations): "Well, this is all wonderful and good; you've presented an exciting future. How do we get there?" In response, at first, I gave all the standard answers: "Find faculty that are doing cool things, and bond with them, show them how the library can help; disseminate news of the world to your colleagues; insert yourself into research and education whenever possible; and be a force for change."

I recognize these are easy advice trinkets to proffer, and they are not sufficient, given the tremendous changes that libraries are obviously capable of perceiving, yet not addressing. There are deeper changes that I have been contemplating as possible responses, and so I eventually relented and spoke them. I fully realize that I am not in the management of a library, and these may not be the most appropriate responses. But I am also certain that the inevitable rejoinders, "These are too difficult, too complex," or "We cannot remove these functions because our faculty continue to expect them" are inadequate. The tiller of change is advancing on the field, the corn has been harvested, and the stalks will soon return to the soil. We better be thinking about the new crop, or the field will lie fallow.

I appreciate the severity of my suggestions. I think this is a conversation that libraries should have, openly and vigorously, because half-measures will not suffice. I may be wrong in my specifics; an out-of-step troubadour with atonal music and lyrics in an offensive language. That's fine. Other and better learned troubadours must surely exist to play the music that will capivate.

Here we go ...

1) Permanently and significantly reduce cataloguing staff. Now. Most cataloguing is superfluous; I am not going to defend the underlying proposition, as there have been numerous assaults already on this point. Original cataloguing - of which there will remain plenty - has an important role moving into the future. Copy cataloguing - the relentless duplication and continual iteration of obscure, underutilized metadata - is absurd. The majority of the useful and attractive metadata is easily obtained through both traditional (CIP) or newer (ONIX) data sources. Don't worry about the rest. There's enough richness in even the essential cataloguing data to do things far more compelling with search than we do now.

There are other costs beyond salaries at account here. One thing that libraries routinely do badly is to get books onto shelves quickly. Every book, seemingly, must be looked at, considered, and metaphorically if not physically measured. Natch. If you keep choose to keep buying books, get them into circulation.

2) One of the key services that libraries provide is the acquisition of content through multi-year licensing contracts. These contracts are ever-more stable and homogenous. Model contracts have been widely disseminated, and efforts like NISO's SERU promise to simplify many routine business arrangements. Although there is continual churn in titles, it is usually not profound, and can be routinized. Where significant contracts must be re-negotiated ... frankly, librarians make poor businessmen and worse negotiators. Acquisitions - the business part of it - should be removed from the libraries and placed into central campus purchasing units. These units, uh, purchase all the time. Move the most critical staff - the ones with the most expert or specialized knowledge of vendors - to central campus units and terminate the remaining positions.

3) Design and implement library systems - industry-wide - that integrate acquisitions with central campus (or city and community, in the case of public libraries) business applications through service oriented architectures or other locally tractable technical means. These high-overhead functions should not be needlessly replicated. Libraries can talk about special needs until they are blue in the face, but there is nothing here that a handful of programmers cannot solve if they are left sufficiently unencumbered.

4) Get ready to stop buying books. I hope that doesn't sound heretical; it shouldn't. We are rapidly advancing toward a future where libraries will be licensing access to books digitally, and they will increasingly be consumed digitally. If there is a perceived need for paper copies, then they can probably be relegated to regional storage facilities. Say what you will about today's superiority in paper feel, but digital has huge advantages in discoverability, access, and consumability.

There is a market here - for a while, at least, until miniaturization yields its inevitable bouquet of optimizations - for libraries to install local equipment for print on demand, like the Espresso machine from On Demand Books; there may be alternative network-resident service offerings in the future. The key thing to consider is diminishing physical inventory, and outsourcing the cost of paper handling to those who need it, and are willing to pay for it.

Libraries should start preparing their campus/city/community purchasing agents: we will have to commence negotiations for large-scale book licenses soon, and maybe not from publishers. Google will be here, maybe Amazon. Think afresh. This transition isn't here yet, but the floodwaters approach. Forget the sandbags. Let's build boats.

5) Tear up the library. And for that matter, tear up the central campus IT organization as well. What we need to develop is something wholly new - an assemblage of semi-autonomous units, each with their own dedicated staff, working in a wholly different kind of campus organization that is neither part of the library nor IT services. These faculty- and student support units should specialize in areas like mobile applications; cyber-infrastructure; web 2.0 apps; rich media for learning; and virtual environment infrastructure. For lack of a better name, let's tentatively name this new organization "Data Services."

I've seen prototypes of this new kind of organization - UC Berkeley has one, for example, although it is still in IS&T. Columbia Libraries has been supporting efforts. But I think they would do better as a separately organized enterprise, formed from staff in the library and IT groups that are laboring to move into the future. Let's extricate the staff from the living-dead prejudices and bureaucracies that confine them, hindering their creativity and hobbling their boldness with the restrictions imposed by organizations possessing conflicting and complex assemblages of aims. Provide them sufficient start-up funding and lay holy hands of prestige upon their heads, but once endowed and blessed, send the children into the world. Libraries and IT groups might well form a temporary joint stock ownership of Data Services until the new unit can IPO, but they must not serve as administration.

New libraries can light the way for new campuses.

Why effect these wrenching changes?

Because It's Time. The librarians that I have talked with - admittedly, they self-select, although they represent a cross section of functions - are ready to move forward into the future. We know how much things have changed; we're not kidding ourselves. We know there should be fundamental transformation, even if some of us necessarily and selfishly want its expression to come in just-a-few-more-years. Yet, a bright future beckons. It is exciting, and we can be part of it. We know it - we have a huge range of skills, and we're bursting with new ideas.

We want to do more. We want to engage. Let's go.