Re-ToolingAcademic Libraries for the Digital Age:
Missions, Collections, Staffing

California Academic & Research Libraries
Third Annual Conference

3d Keynote Speaker:

Dr. Clifford Lynch, Director,Division of Library Automation,
Office of the President, University of California

"The Changing Nature of Collections in the Digital Age"

[The following text is a transcription of a talk given at the Third Annual Conference of CARL (California Academic and Research Libraries), in San Francisco, CA, on Saturday, October 21, 1995. Bracketed words and phrases are inserted by editor.]

Good afternoon.

I thought it would be interesting today to spend about 45 minutes or so "romping through" the changing nature of collections and the issues around collections.

This is something that I think all of us are grappling with in different ways. It's an increasingly pervasive issue as we try to figure out what libraries will be, and how they will inter-relate in the coming years.

When I started making notes for this talk a couple of days ago, I assumed I would be able to paint a coherent picture. But the whole issue of collections and contents is just fragmenting into many different pieces, each with its own intrinsic rules, and trying to come up with anything that spans across these coherently turned out to be very difficult.

So I thought I would focus on four or five different areas of interest, talk about the issues within each, and reflect on the very deep divides between them.


The first area, of course, is traditional print collections. These are, unfortunately, or fortunately--depending on your viewpoint--not going away. It is a little scary how many people seem to BELIEVE they are going away, faster than anyone could imagine possible. It's interesting to talk to people--including some university administrators as well as folks in the corporate world--who don't deal with libraries very much, but have strong opinions about them. When they ask me what I do and I explain that I've been working in library automation for many years, they reply, "The whole thing's electronic now, isn't it?"

You have to contrast that perception with the reality. I've been informally polling library acquisitions for the last year or so (and this is one set of statistics that I wish ARL would track). A typical library is designating less than five percent of its annual acquisitions budget to electronic content. This is obviously NOT an overnight mass migration to electronic content.

[Access versus ownership]

The major shift in collections acquisitions philosophy, other than coping with prices going up and numbers of subscriptions going down, can be described with the slogan (coined in the late 80's and early 90's) "access as opposed to ownership" or "just in time, rather than just in case." This trend is realistic, when you consider the difficult financial environment. Basically, you want to buy little pieces of things, rather than large aggregations of materials that you can't afford and that are lightly used. There are a couple of aspects, though, that people don't like to talk about, but which I will discuss briefly.

[Changes in relationships among libraries]

The first issue is that, if you want to go for access, you're assuming that somebody somewhere owns what you want, and that you can simply locate the material and get it. Libraries would evolve into two groups: one group would hold relatively limited local collections and acquire on demand for its user community from other sources; the second group would, among other things, continue to amass a collection and makes it available to all of these access point libraries. The implication of this arrangement is that the relationship between the two types of libraries would become more and more financial and contractual as time went on. Back in the days when there was more balance in the system, you could handle collection sharing through the traditional interlibrary loan arrangements. But as the two types of libraries become more and more distinctly defined, some very tense contractual relationships will basically formalize the fact that some libraries are relying on access to other libraries' collections. There are already signs of commercial players becoming involved, right alongside the "collecting" libraries, and these commercial players may become the supply points to libraries which just want access. I think that the access versus ownership approach is a very interesting local tactic, especially for the "budget-constrained" library; but I think that we have been very cautious and reluctant to think through the systemic implications of that for the entire library system.

[Monographs at risk]

The use of "access versus ownership"--at least as I have seen it played out--has been primarily with periodicals. There is a gap between "acquisition by the article" and "acquisition by subscription." But monographs seem to be left behind. I am beginning to believe that monographs may become the mid-range casualties of the migration to electronic delivery. Journals are pretty easy to move by fax, and pretty easy to print on demand. They have economic properties around them. They are aggregations of articles and often the unit of real use is the article. The monograph, on the other hand, tends to be difficult to navigate in electronic form. It's not something you want to print on demand, and not something you want to move by fax--at least today. I keep hearing that the monograph is more and more an endangered species. Scholars in fields who rely on them feel that the various economic and technology pressures here are creating a major problem for monographs.

[Importance of abstracting & indexing databases]

An unexpected byproduct of this "access versus ownership" trend has been a much greater emphasis (than one would have thought ten years ago) on the importance of abstracting and indexing databases. If you don't have the material, it's increasingly vital to provide access to bibliographies of what's out there that you can acquire on demand for your user community. And the availability of decent access to abstracting and indexing databases in various fields has become one of the essential factors that has made this approach popular. But I don't think that some of the accompanying dangers and strange byproducts of this approach are sufficiently recognized. For users who no longer have material locally to browse, but can order items on demand, what's offered in the abstracting and indexing databases essentially defines the literature. What happens to all the materials that are not included in the indexing and abstracting databases? We are going to need a serious, systematic analysis of the gaps in abstracting and indexing services, in view of the scope of scholarship and the scope of material to which research libraries want to provide access. And that's a project that's will become increasingly urgent over the next few years, as more and more libraries are forced by economic pressures to move to access models.


I'd like now to turn to electronic print--that's a nice oxymoron, but I don't know what else to call it. What I want to convey is the notion of traditional publishing as it is carried out by electronic means, and where the players are still the traditional publishers. Now, had you asked me ten years ago to speculate about the future, and how electronic versions of what started life as print publications might evolve, I would have predicted a much more incremental and smooth transition. There's a natural extension from document delivery models that deliver fax to models that let you print copies of materials from some electronic database. But I think what we're finding--and this is partially a function of various economic and legal issues--is that there is an ENORMOUS chasm between our practice in physical print and our practice in electronic print, as it is presently evolving. There are completely different rules for this electronic world, and the gap here is much wider than I would have predicted.

[Costs requiring new contractual arrangements, aggregated buying and selling]

There are a couple of the issues surfacing. First of all, because we're operating in an environment of license, rather than purchase, transaction costs are extremely high. Negotiating a contract is expensive, compared to writing a check for a subscription that's simply covered by the normal processes of purchase. That, in turn, is leading to a strong incentive to do bulk purchases. There's a lot of overhead in the way we make case-by-case selections about print, and when you add on the burden of contracting and figuring out how to mount and deliver electronic sources for a user community, there is, I believe, a very strong economic and operational trend towards aggregating multiple publications from a single publisher or even materials from multiple publishers under a single contract. I think we're seeing that both in the way that third party aggregators--people like UMI or IAC--are finding a niche aggregating large collections of material from multiple publishers and offering it under a single license through a uniform delivery mechanism, and also through some of the rumblings we're starting to hear from the big publishers: "Well, let's not talk subscription by subscription, necessarily, let's talk about our 'relationship' with you as a supplier of content." I think that will be an inevitable consequence of the whole licensing and delivery framework for electronic material.

[Direct publisher-end-user relationships]

There are also some other issues around publisher relations that are really scary. The publishers are, I think, trying to figure out whether they are going to be direct content providers or work through third parties. The prospects here for balkanization are fascinating. Talking to typical working scholars, I haven't found anyone who says to me, "Well, let me tell you about my important research readings. They're called Springer-Verlag." They still tend to characterize them as subjects ("I do research in differential equations," or "in 17th Century French history"), not in terms of material published by a particular publisher. Yet we're starting to see publishers set up direct delivery sites, each one different. We're actually starting to see bizarre notions surface: Publishers--because they'd like to get right to their end-users, their end readers--would like to authenticate them individually, by giving them ID's and passwords. I began to visualize researchers with huge cribsheets pasted next to their workstations, with ID's and passwords for every publisher that produces anything in which they're interested. And of course there's a different user interface on each one of these--that's just assumed here. [laughter] This is a very different world from the real coherence that libraries have been able to supply in print, whether it's in an access model or an ownership model. There's a certain coherence and uniformity to print that I think is in terrible danger of breaking down here.

[Archival responsibility]

The other big, unknown issue in this world of electronic print is archival responsibility. It's sort of understood in the physical print world that the archival job rests with the libraries, primarily, and not with the publisher. But it's a little bit ambiguous, in the world of electronic publishing, who holds that archival responsibility. Indeed, it's a little ambiguous whether, in fact, libraries have the requisite permissions and legal capabilities to take on an archival role. I'll just direct your attention to two interesting developments in this area, which underscore the unsettled nature of the assignment of this archival responsibility.

The first is the Lehman white paper, proposing revisions to the intellectual property laws in the NII environment. It contains some language that is proposed for inclusion in the copyright law amendments having to do with the use of digital techniques and the ability to digitize for archival purposes. The second--which is perhaps more graphic--was a Task Force that was jointly sponsored by the Commission on Preservation and Access and RLG, and chaired by Don Waters, from Yale University, and John Garrett, from the Corporation for National Research Initiatives. The Task Force was looking primarily at preservation issues around "indigenously digital" content--items that weren't just scanned images of paper, but started life in digital form--and it actually went so far as to suggest that we what we may need to create are "store-only" archives, that would maintain copyrighted materials in digital form. These archives would have legal permission to store the material, but they would have no legal permission to provide researchers access to it, until after copyright expired and the material moved into the public domain. But at least these sites could hold the material in trust until that time, rather than letting it all evaporate.

It is interesting that this group even proposed such an idea, which is bizarre, but interesting, and perhaps pragmatic, in the sense that you might be able to get it through legally. This is one of the few scenarios that you might be able to get the rights-holders to accept. It's kind of a measure of how desperate this concern about right-of-way for the archiving function really is. And I think that this issue of archiving responsibility in these relationships is a very critical one.


Now, let me turn from this area of print publication by electronic means to talk about the Internet, and all the stuff you can find there, because this is another area where, I think, every library is struggling: "What does the Internet have to do with collection development policies?" "What does it have to do with our approaches to bibliographic control?" "We don't know what to do about this thing."

[Internet NOT a library]

Let me make a couple of suggestions. I would argue that the Internet is couple of things. It is a way to reach electronic information services that someone has constructed. It is a publication or information dissemination channel, a way that people can make things public, if you will, very much like running a printing press or a TV station, only with a much smaller front-load investment. One of the things that I DON'T believe the Internet is--contrary to certain popular sloganeering--is a library. This is not an organized collection of anything. Rather, the Internet contains the same sort of collection you'd get if you just said, "I'll put out a big hopper and collect all the output of all the publishers, all the Kinko shops, and all the TV channels." The result is material that can rationally be drawn upon to build collections, and to be part of collections; but as a unit it doesn't have any single organizing motivation, in the same sense that collections do, as we think about them in the context of libraries.

[User behavior: the attraction of instant gratification]

The Internet is a source for material. But it is a very strange and ambiguous place when we start looking at it as a source for content to add to library collections. It's strange and ambiguous for a couple of reasons. One is that it's a world of immediate gratification to users. Consider this: I think all of us have seen tools on the Internet like Lycos and Infoseek that have a sort of superficial similarity to a really BAD online catalog. [laughter] They have one thing that your online catalog doesn't have. It's called an instant gratification button. It's that underlying element of, "Oh, this looks interesting. Let's go have a look at it right now." So, it's an extension of a phenomenon that we've been seeing signs of for some years, since we've started making electronic text available.

A little anecdotal material from the University of California: We have a number of abstracting and indexing databases on our online system. Some of these databases also incorporate full-text components. However, the full text is not comprehensive. In other words, you will have an abstracting and indexing database that attempts to cover a discipline or a class of material coherently, and attached to it, fairly randomly, you'll have full text of maybe a third, a quarter, or a half--probably not more than a half--of the articles in there. And that's because the supplier of this material couldn't clear rights for the other half, or couldn't come to agreeable contractual terms on those rights.

Now, it's interesting to see students deal with this. There is at least some strong anecdotal evidence that students, once they discover the command "Restrict my result only to those things that are available in electronic form, and don't waste my time with all the other stuff, because it's 11 o'clock and the library's closed, and my paper is due in the morning," succumb to immediate gratification. Being able to read documents in electronic form is very compelling, particularly for people who are in a hurry or want an answer and aren't necessarily doing the most rigorous imaginable scholarship. [Laughter] (None of us would ever do that, right?)

Well, I think you can view the Web, as a user sees it through these kinds of finding aids, as the extreme extension of that philosophy. These catalogs--and I use the term advisedly--of things on the Web are really lists of everything you can have right away, for the expense of a mouseclick or two.

Users are tending to bypass library catalogs in some cases and to look for answers on the Web because that immediate gratification is so compelling. I'm sure every one of us has stories about people who spent four hours surfing the Web looking for an answer that they could get from an encyclopedia in two minutes. It's just that the encyclopedia wasn't on the Web, or they didn't have access to it, and it's so much fun to sit there and get what you want.

So we have this kind of strange disintermediation showing up where there is a growing class of users who tend to bypass libraries altogether for this sort of immediate electronic information gratification, and I don't think we've done a particularly good job of trying to integrate our own systems that provide access to library collections, and that provide, to the extent possible, access to content, with this kind of mind-set. They're really presented to the user as different worlds: Here's the world of the Web and instant gratification; and here's the library catalog, where maybe you can get some of these things quickly, and maybe you can't. It's a very confusing issue for users, and I think we're seeing users choosing the Web over automated library catalogs, just because of that gratification.

[Integrating Internet resources into the library "collection"]

As we struggle about what to do with information on the Net and how they relate to library collections, we need to consider the tactics of how one might incorporate that information into a collection, and how one might select and acquire that information. We can't quite figure out whether we're comfortable putting in pointers, or whether we want to copy the item and take it under the sort of operational umbrella of a given library. Putting it under the operational umbrella of a given library by making a copy is attractive, in that it gives the library some handle over archival issues and over continuity of access issues. As I think we all know, continuity of access is kind of a shaky concept for certain material on the Web. But here, we rapidly encounter again this sort of quicksand of ambiguity about intellectual property. If you stop and think about this, it's really scary.

There's this sort of notion that, if you look at a piece of print, you don't have to make a copy of it to look at it, and in particular, to look at it and see the copyright notice on it. Simply pulling something off the Web into your browser, before you've even read the copyright statement that says, "You can't do that," is considered--at least in some people's view--making a copy. Of course, I think you can make a counter-argument that, "Well, they didn't just put it out there, probably, to entrap you." (They wanted you to look at it, maybe.) But this illustrates how really shaky this ground is, and when you go further, saying, "Well, I think I'll just copy these documents over and put them in my own local collection so that I can manage them on an ongoing basis," it becomes more ambiguous still. The notion, however, of first trying to identify the rights-holders for much of the material on the Web, and then find them, and then get their permission to move copies of this material into a local collection, makes the kind of contractual issues around dealing with print publishers who are supplying materials in electronic form now, attractive by comparison. So I think we face some real problems with documents on the Net, and how to integrate them into collections, and I think that we really are going to have to focus on narrowing the gap between material on the Web and material in collections so that users see this in a more coherent way--so that users are no longer seeing these as fundamentally disparate spaces of information. I think we're going to have to deal with that, and that's going to mean moving catalogs and other bibliographic tools into closer conformance and closer alliance with some of the things that let you navigate the Web.

[Semi-private information domains]

I think that another symptom we're seeing here is that the Web and related tools make it very easy for people to have private or semi-private collections of information that they're sharing among a group of colleagues, and to allow navigation and access to those in a way that's consistent with the way they navigate public space. We've never done that well with libraries. Libraries are one place, and personal and group information spaces are another, infinitely distant and separate place. I think that the networked information environment of the Web has broken down in a way that users view as desirable and advantageous. And I think we need to think through how to make libraries' collections fit into that kind of a model as well.

[Documents versus information services on the Net]

So that's the world of documents on the Net. Now one of the other problems with the Net is that it's a zoo of virtually incommensurate objects. My favorite example of the problems we encounter is when someone does an Internet search and retrieves three items--two documents and the Library of Congress. [laughter] There's this sort of mismatch between things of very small granularity and enormous services like NEXIS or the digitized special collections of the Library of Congress, which really form their own kind of information space, with their own navigational tools and their own classification schemes, and their own internal logic. Yet we find in the networked environment these paths to big information spaces intermixed with quasi-document items that are scattered around on the Web. We don't know what to do with these services. We've never really had a good handle on how to fit a service, which might hold inside it as much information as a good-sized library, into the holdings of another library. How do you meaningfully catalog LEXIS or NEXIS so that a user sees it in a way that's commensurate with seeing individual books and journal subscriptions?

This is a concept I think we've never really been very comfortable with, and it's always been a little awkward. But it's going to get VERY awkward, I would argue, because more and more of these digital libraries are coming on stream. And I think one useful way to think about digital libraries (and I'm not even going to get into the argument about whether digital libraries is a good term or not) is that they're sort of organized information collections and access services that are accessible through the Net. So I think the issue of describing these effectively is going to become EXTREMELY important four or five years out, as digital libraries continue to proliferate.


So that's a few words about the Internet as a third source, alongside traditional print and electronic publications, as resources which can comprise the collections of the future. Now, I want to touch on a couple of other areas. I also want to finish early enough so that we have some time for questions and discussion. One thing I don't want to say very much about, but I would be remiss not to mention, is special collections. These are unique collections of materials that individual libraries hold, and of which these individual libraries are now in effect becoming publishers. Basically, libraries are starting to open up their special collection rooms into published items on the Net, and they're doing this sort of half under the rubric of access and half under the rubric of preservation and protection of fragile and specialized materials. It's clear that those are going to be important offerings from many research libraries on the Net.

[The problem of priorities]

There are a couple of hard problems there--difficult management and technical challenges. The first is what priority to assign to special collections. How do you balance investing in your responsibilities or opportunities as a publisher to the rest of the community with your desire to spend your investments on acquiring content for your user community?

Most special collections are, by their nature, of interest only to a relatively narrow range of scholars. That's not always true, but certainly there are many cases where that is largely true, or has been in the past. So I think that how to balance those investment priorities is a hard challenge. I think it's interesting to note how much of the work in digitizing special collections seems to be an opportunity for finding one-time grant funds and things like that--funds that are outside the normal operating and acquisitions budget. I worry a great deal about what could happen if some of those funding sources dry up--given what is happening in Washington.

[Transcending geography: "logical" special collections]

There's a second challenge about special collections, which is more intellectual and more technical. Special collections have always had a weird geographic aspect. For example, you could locate a classic archival collection of the papers of some personage. But then you might also find there are other libraries that have this person's correspondence, or correspondence from other people who worked closely with this person. Historically, it's been up to the scholar who is mining these archives to write books, or write articles, based on examining multiple collections, that bring these materials together and form links. Now, there is a new option emerging, which is the option for the keeper of a special collection, perhaps working in conjunction with groups of scholars, to use these special collections as material for making more coherent collections, because you can now transcend geography and the accident that some of this material is here and some of this material is there. And I don't think we've really thought through the ramifications of that yet, and particularly who might take the lead in building and underwriting such "logical special collections" (if you will) that are built up from geographically disparate and now digitized physical special collections. I think this is an enormous opportunity, but one we're still trying to understand how to exploit.


Now, another area I just want to mention--which will lead me into the final area I want to discuss--which is clearly on the minds of people trying to understand the scope of collections, is various kinds of sensor and digital data. These are things that are not (if you will) authored in the traditional sense of someone intellectually performing an "act of creation." These are collections of data that are gathered from satellites, or from various kinds of telemetry devices, or from tracking transactions in a market, or something like that. These come in two flavors: One is the sort of passive data set, and the other is the active information flow. Passive data sets would be, for example, maps that have turned into digital data sets in the past ten years, or the collected readings from various remote sensing platforms (two decades of snow measurement in the mountains of the Sierras).

[Use entirely dependent on expensive software and computing capacity]

Now, these data sets are a big problem, because--back when this was paper, we could sort of deal with it by benign neglect. You could read a map like you could read any other paper--you didn't do a lot of training in map-reading. Now, all of a sudden, these common items are uninterpretable without software. And the problem is that the software fills a huge range. At the low end, it may be a pretty dumb viewer--something that just lets you look at a digital map and scroll around, and maybe punch in a street name to bring up something. At the high end, there are analysis tools that will consume a super-computer, and fifteen graduate students, and all the things that come with that, in order to make good use of these tools. And the library faces the question of whether it should do a little more than just keep the bits on the shelf. But how far do you go into the software? And particularly, when you recognize that there is a lot of leading edge research in many of the fields that use this content, what can you do with that software, how can you interpret it, and what are the new methods for applying that software? What's the training role? So here's an area again where I think roles are very ambiguous and the departure from paper has really emphasized that ambiguity. In a sense, it's perhaps made it clearer, and, in the long run, sorting this out may be really beneficial, because it will get us away from the huge archives of printed numeric weather readings and such that are collecting dust and filling up buildings.

[Data flows]

The other side is even more interesting, I think. And that is the idea of data flows and sensor flows. We tend to think of data sets as passive, and something that weighs on me increasingly in the last year or two is the fact that we are moving, very quietly but in a very real way, as a society, to a place which is full of continuing real-time data flows. Some of it is just being blasted out, because it's fun; some of it is being sold; some of it is very valuable, and in some cases, you pay a lot more to get it in real-time than you do to get it a little bit later than real-time. This ranges all the way from seismic sensors to the radio channel that tells you what time it is, in digital form, so you can set your clock to a couple of nanoseconds off of world-standard time (if you're worried about that), [laughter] all the way through to great security cameras of Los Angeles, that let you look into every Seven-Eleven for two minutes every night. You're starting to see areas of cities, now, where they're putting up video surveillance, just as a way of trying to keep down street crime. You have all of the financial "telemetry" that flow out continuously--all of the equity, bond, and currency market transactions. You have real-time space imagery that is being marketed. You have a huge amount of this kind of information showing up. You have all the news wires. Now, there's this real ambiguity about what you actually do with this stuff. Is this something that's part of a library collection? It's not static--it's a flow. Is providing access to a flowing river a library function?

And it is an IMPORTANT--it's an INCREASINGLY important--information source in some areas. Furthermore, there's the question, "Does anybody record this stuff?" Because each flow, of course, has as its dual a database, which is the archive of the flow, and in some cases, this archival database gets rather large. Just talk to the people who worry about where to put all the satellite-based remote sensing data, who are planning systems like EOSDIS, which is talking about terabytes a week. The archives on these things are ENORMOUS, in many cases. It's not at all clear to me how those fit in collections, but it is clear to me that these are important enough to some people that they pay for them, and in some cases, the opportunities offered by these archives may be something that our next generation of scholars will spend quite a bit of time exploring.


Now, that takes me to the final area that I want to mention. We've talked about some of the offerings on the Internet--items that are sort of like documents, things that are sort of like electronic information services. Well, there are also appearing collaboration environments that are discussion environments. Some of them are as trivial as listservs. Listservs are a very trivial thing, in a way. They're not very high tech. They just redistribute and perhaps archive ASCII text messages. But it's striking the extent to which these things are starting to structure communication among groups of people in some communities. The impact of this notion of mailing lists, and being able to set them up easily, and let people come and go from them, I think is going to turn out to have been a much under-estimated phenomenon of the early 90's. And these are the simple case.

What we can see coming already are distributed collaboration environments that let people share text and video, and talk to each other, and operate scientific apparatus, and review data. There are already some interesting prototypes. And they leave trails. Sometimes people replay sessions that occurred in these environments, because these acts of collaboration and communication are recorded. That's one of the primary properties of the digital age: Use generates a recording of that use as a byproduct. And we don't know what to do with these recordings--or, who should take charge of them, or whether they're even important, or how to find anything in them. One of the really "evil" properties of video, for example, is that although we're really good at flipping print pages, trying to find something, we really haven't done all that well in developing technology, yet, that lets you go through 40 hours of video in five minutes looking for the part you want. I would speculate that that's one of the reasons why things like taped lectures are so awkward for many people. It's very hard to get at particular pieces within them. But we're creating essentially this kind of collection of recorded discourse--among scholars, among all sorts of peole, knowledgeable or not, with the random spammer jumping in for flavor. There's a very real question about what role libraries should have in incorporating this kind of material into their collections.

I would be remiss not to note that copyright in this area probably makes some of the intellectual property issues around the Web look simple by comparison. And I think the best way to describe the intellectual property issues here is "Nobody's quite sure about who has rights to what."

But we're confronted with this growing corpus of material that may or may not be of lasting value, and we're not sure how to describe it, how to save it, how to navigate it, or who will use it. I guess I would suggest that, over the last few decades, libraries have been relatively resilient in avoiding being over-impacted by new forms of media. [laughter] Just as we're talking about digital materials down at the five percent level, if you look at video materials, audio materials, for most libraries (I've never done the survey there), I'll bet they're not at the fifty percent level either. Yet we're in a society where people do a lot more information collecting from TV than they do from reading, in many parts of society. I guess I wonder a lot about the extent to which libraries really will be comfortable and eager to deal with this side of the changing collection opportunity, or challenge, or issue. And to what extent--because this material is strange, and we don't know the rules about it yet--it will just be marginalized in much the same way that I think a lot of video material is marginalized. Of course, we also will have to understand the rules of ownership and use on this material.


[Where we stand]

Now, just to bring this full circle, I want to come back and reflect on these different classes of material, and the amazing variation between them. We have traditional print--we understand how that works. The economic, and legal, and management framework for that is pretty solid. We have this business of electronic print--traditional publishers going electronic. I think here the legal framework here is relatively clear--it's one of license, rather than purchase; but we've come up with a modus vivendi here, where we can operate. What's striking is that the rules are very different, the practices are very different, and the tactics are very different. We have the world of things on the Net. There, we don't understand the rules very well. And our users view that in a very disconnected way, I think. Sometimes, we view it as a very disconnected thing--something that's outside of libraries, but is a useful alternative resource. I think that's an area where we haven't come to terms. Then we have special collections. Again, here, I think we understand the rules. What we don't understand so much here is the priorities. And we've not fully absorbed the implications around the dissolution of geography in the electronic environment, as they relate to special collections. We have the whole matter of data, as distinct from intellectual works--and I know that's a sloppy distinction. Here we have real questions, I think, about library roles, and perhaps more than roles, as we'd like them to be: What's do-able? What can we really afford? What have we really got the capabilities to handle?

I think, in the area of data, we also have unprecedented opportunities to forge closer cooperations with working communities of scholars, because I think, ultimately, we will recognize that management is going to be a joint responsibility and a joint enterprise. And, finally, we have these emergent forms of scholarly communication--these new media objects, if you will, some of which are being created by a function of our increasing ability to record everything in sight, and our tendency to figure that "Oh, if we can record it, we probably should!" Think about some of the discussions about instructional use of technology. One by-product of this is that it's not at all implausible to think about people recording every lecture given in every major university every year. One can, of course, ask "How many recorded Calculus 1 courses do you really need--or want?" But we're moving into this kind of environment. And here, I think, is the place where we probably face the hardest problem, because it's more than a problem of tactics. I think, to a great extent, when we look at the problem of how to acquire material on the Internet and what that has to do with our collections, it's certainly a tactical problem of "Do we move it inside?" Do we store a pointer? How do we move it? Can we move it, legally?"

But I think that, when we look at the new forms of scholarly communication, it's a more fundamental problem of making assessments about how important this material is, and how much of a priority it should be for libraries, and what libraries should be doing with it. And part of it, also, is coming to a broader understanding, as we look at the user community for this material, present and future, about what they need from it.

[Loss of coherence]

I think that, as we look over these various aspects of collections, it's amazing to me to see how much of a loss of coherence there really is among them: how it's very easy to specialize in one, and focus in one, and make progress in one, and because you're making progress, and you understand it, and it's easy, to ignore some of the other collections. I think that we will see some institutions do exactly that: focus on one or two of these areas. Maybe that's not a bad thing. Focusing on all of them at once I think is very hard, very challenging.

[Where to assign resources]

Having mapped out these areas, I'll just finish by pointing out that we're also faced with the problem of assigning resources among them. It's not just a question of what do we do in these areas--it's how we assign our resources from one area to another. Now I think that, today, we're still dominantly in the traditional print world, as far as our allocation of resources. I think that it would be very interesting to come up with some taxonomy of non-print areas. Maybe the one I've sketched out is the right one, maybe it's the wrong one. But to track seriously how resources are re-deployed over the next decade or so into these areas, and to see whether these really continue to be sort of marginal areas--issues that are exciting to talk about, because they're different, but that, ultimately, from an operational basis, are kind of marginal--this is an important measure of progress. Tracking them is going to be hard. For example, if you look at the Internet area, one of the most disconcerting things about thinking about acquiring material from the Internet is that you don't have to pay for most of it. I mean, what kind of "acquisition" is this? You didn't have to pay for it! That immediately means we don't know what to do with it, how to assign value to it, how to include it in our ARL statistics.... [laughter] And I think that there will continue to be some material out there for which we don't have to pay, that are just part of the circle of gifts. But we will pay for supporting them, bringing them in house, managing them, and providing pointers to them. So really getting a handle on how much resource is going into these different areas is a much more complex analysis, I would suggest, than just looking at where the acquisitions budget goes. I think it'll be very interesting to see how the proportions, and the emphases, and the resource commitments, shift--or don't shift--over the next ten years. We will want to track this.

But I hope I've at least given you a view of the terrain, and maybe some of the ways in which it could shift.

Welcome to the wonderful and very fragmented world of the future of library collections in a networked environment.

Thank you.

Return to top | Return to Conference Program