Eugene, Oregon
May 13-15, 1998

The Future Network: Transforming Learning and Scholarship

Program Session IV

Research Library Leadership: Network Policy and Applications Development

A Facilitated Discussion

James G. Neal, ARL President, Convener
Johns Hopkins University
Clifford Lynch, Discussion Facilitator, Executive Director
Coalition for Networked Information

MR. NEAL: The presentations we have already heard today offer a framework for this final program session, one in which we participate and one that will, I hope, prove to be a very stimulating and exciting discussion of research library applications of the future network and of how we as librarians can play a very effective role in network policy development.

About a year ago we had a similar discussion, facilitated by Doug Bennett, that helped us think about the future of scholarly publishing under the impact of technology and under the impact of some of the difficult budget and pricing situations we face. Out of that discussion came SPARC, which we are now developing in very aggressive ways. I hope that same type of discussion can take place today and that we can create a framework for our ongoing participation and leadership in developing the future network.

I can think of no one better qualified to facilitate this discussion for us than Clifford Lynch. As the Executive Director of the Coalition for Networked Information (CNI), Cliff is focused on the agenda for CNI’s 200-plus organizational members who are concerned with information technology, networked information, and the role these play in scholarship, learning, and the overall area of intellectual productivity.

Cliff has a Ph.D in Computer Science from the University of California-Berkeley, is an adjunct professor at UC-Berkeley’s School of Information Management Systems, and is certainly a friend of the academic library community. Please welcome Clifford Lynch.

MR. LYNCH: I will outline some of the common themes we heard from other speakers today and also try to present a couple of issues that didn’t come up that may be helpful in thinking about these themes. I will then open the floor for discussion by presenting a few specific questions to frame our discussion.

We really heard three challenges today. Two of them were pretty explicit and one was more implicit. The first challenge we heard was very direct, that is, that the higher education community, in conjunction with a group of industrial partners, has launched on this Internet2 enterprise and that enterprise is in fact being supported by many components of the federal Next Generation Internet (NGI) funding program. This is all about applications. It is particularly about getting some experience with the development and moderate scale deployment of a whole new generation of network applications that let us reach way beyond what we can do practically with today’s Internet. Hopefully that will let us realize, on at least a modest deployed scale, some of the vision that has been incubating in labs and in other very constrained kinds of experimental environments, or perhaps existing only in the minds of scholars, librarians, researchers, and educators over the past decade or more.

It may be helpful to remind you of two specific things about Internet2 to keep in mind as we discuss it: Internet2 is really not, at least as I understand it, intended to be a substitute for the current Internet. The current Internet, as you know, still suffers from reliability and capacity problems, even in doing what it was designed to do today. Those of you who have, for example, committed your institutions to models that involve remote access to commercially-offered information resources find yourself relying every day on the network working correctly, and, as we all know, some days it doesn’t do so well. Internet2 is not intended to be a fix for that. In fact, there are acceptable use policies that, at least as I understand it, will be put in place on Internet2 that essentially bar transit traffic. So, for instance, you can’t hop to the Internet2 site nearest your favorite commercial provider and then take that final hop onto the commercial Internet. That’s not what Internet2 was designed to do. I do believe that if there are very specific collaborations involving content providers in high-performance advanced applications, something might be worked out to do those kinds of experimental or exploratory applications in the Internet2 environment. But it isn’t an appropriate way to patch over the current kind of production problems with the Internet. That’s important to remember.

The other thing that I want to remind everybody about—because it’s very important as we think about library strategies with regard to Internet2—is that Internet2 begins at home. You heard all of these discussions about the enormously high-speed backbones. What that boils down to is that somebody will drop a very fast pipe on the edge of your campus if you’re an Internet2 member. The specific applications that you will need to work on Internet2, and to use that kind of capacity, need to reach that pipe across the campus. If your campus is currently wired with something like a ten megabyte per second switch, you have a serious impediment to this match. In many cases I suspect that institutions are looking to applications development funding that they may receive on their campus via routes like NSF to build up local capacity, for example, between that drop-off point on the edge of the campus and some particular group of labs or other facilities that are going to be running these Internet2 applications. If the library is going to be a major content supplier and needs those kind of Internet2 capacities, then one of the pieces of strategy needs to be working within your own campus environment to make sure that you have the connectivity to make use of the Internet2. That is the first challenge we heard today. Internet2 is the unprecedented opportunity to explore new kinds of applications we couldn’t do before.

What kind of applications can libraries play a role in that improve scholarship, teaching and learning, and access to information? There are a couple of easy answers. For example, one can argue that Internet2 makes the world a lot safer for multimedia digital libraries that rely heavily on delivering video content and things of that nature, things which you really can’t do very satisfactorily over the current Internet. But I believe there’s more to it than just a faster version of what we have today. This challenges all of us to think hard.

The second set of challenges that we heard were about distance education and about how libraries—particularly as they accumulate and organize more digital content—can effectively not merely support distance education, but become a full-scale partner in planning and execution. That clearly intertwines with Internet2 in the sense that Internet2 will be what makes it possible to do considerably more with distance education than we can today.

The third set of issues we heard were more implicit. The way I summarize those is that the network environment—in particular the prospects in the Internet2, but, indeed, even the networks we have today—are changing in fundamental ways, ways that effect the practice of scholarship as well as the artifacts that document and further scholarship and that serve as a focus for scholarly dialogue in teaching and learning.

We’ve heard from a whole range of scholarly disciplines about multimedia, about very complex, layered documentary sites, about dataset-intensive sorts of things where we’re really moving beyond the notion of the heap of datasets into captured virtual realities. For instance, David Schade’s mention of a virtual sky is an example that I find very powerful in conveying a profound transformation.

I think there is an implied challenge to us there. Again, it is very much against the backdrop of Internet2 and NGI: What is the libraries’ position with regard to these changes in the way the scholarship is conducted, documented, and passed along? What will the library’s role be in facilitating the creation and organization of this material, ensuring access to it, and ultimately in preserving it?

One of the things I always find myself wondering as we see these bursts of creativity from various scholars is what will happen when those folks retire. We don’t want those resources to just vanish.

Those are some of the challenges that have been put before us today, and they’re difficult ones. They’re challenges that get seriously at the question of what role libraries will play in the future networked information environment and, by extension, what role ARL might play as an instrument for the research libraries to act collectively.

I’ll just make one other final comment before opening up the discussion. There are many ways that libraries can get involved or that libraries collectively, through an instrument like ARL, can get involved. One way is through policy. I think there’s a sense, though, that the policy, particularly federal policy, is in a lull right now. There was a tremendous amount of work getting the NGI and Internet2 launched. Now it’s time to focus on implementation and on applications. But there are many other policy dimensions that show up, including issues of privacy, of institutional roles, and of preservation.

So when we think about policy in this context we need to think broadly and not merely about interacting with federal policy. There is a role for libraries in leading a dialogue with regard to institutional policy, agency policy for applications and content, and how those things need to be rethought and reconfigured, and funded, in the network environment.

There is a wide range of potential scope or action. Some of it, though, has a little bit different flavor to the sort of policy initiatives that have historically been of interest. Let’s now open the floor to discussion. Let me open the discussion with some questions. What are the applications we should be building? What are the roles we should be playing in this network world? How do research libraries respond to the Internet challenge?

MR. BILLINGS (University of Texas): To me it feels like the three challenges that you mentioned, in my mind at least, raise a bigger challenge for all of us. Basically, that challenge is to discover how we can take advantage of the new opportunities we have with the faster networks, the ability to affect distance education, and the opportunity to help change some of the fundamental ways in which education works and in which libraries work.

It looks to me that what we’re trying to do is—and this is the way we typically solve problems—is look at how we’ve done things in the past. It bothers me that at this point we’re looking at replicating what we’ve been doing in a different mode rather than searching for really visionary ways of solving some of the problems that we’ve had and taking advantage of what clearly will be new and improved systems. These systems are so imminent that if we don’t look right now at the very fore edge we will instead be still spending time working on replicating past ways of doing things.

I may or may not say something tomorrow when we get to the business session about SPARC. I put my money in and I’ll put my mouth behind my support for SPARC at this point, but it seems to me what we’re talking about there is looking at recreating with the tools we have at hand journals as we’ve known them in the past instead of thinking about taking advantage of the great new speeds we will have, looking at some economies in education, and so on. Perhaps we ought to be thinking about simply unbundling journals and taking advantage of the network speed by looking at different policies with respect to the impact that publication in traditional journals has on promotion and tenure in institutions.

You know, we’re all talking on the edges of these things at the present time. But we’re not talking about these things in terms of grabbing the immediacy of the greater speeds and capabilities we will have. It bothers me to think that a couple years down the road we will be thinking again about how well SPARC, for instance, is working.

Somebody ought to be brave enough to really grab a completely new approach to utilizing some of these advances and try something different instead of just slightly altering something that we have already been doing.

MS. MARTIN (Georgetown University): I agree. This may be totally ludicrous, but I watch Star Trek all the time and over the last few years it has suddenly occurred to me that what I’ve been watching isn’t that much science fiction anymore. There is a lot that is within the realm of our technological capacity now. Can we create a vision that goes beyond a five- or ten-year span? If we envision something new, we are more likely to be able to create it rather than have it created for us. So, if we can create a vision that goes at least one generation into the future, putting together our best thoughts about what will happen with scholarly communication, with libraries, and with information, that will then allow us to take the first steps—this year’s and next year’s steps—towards that vision.

But the issue with that is we have to bring our communities along with us. And, Harold, I agree with your comments but the problem with going to a completely different model is you have to persuade people along the way. So maybe we have to go parallel to the past, develop the vision, start taking steps in that direction and at the same time support what we’re doing now and what our users are used to doing and probably feel as though they need to be doing for the near future.

MR. LYNCH: Let me add a couple of comments to this and sort out a few issues. One of the real dilemmas every research library faces today is they’re being told simultaneously to live in the present and to live in the future. So, though there are some significant crises right now having to do with the cost of the kinds of primarily printed materials that libraries have always acquired, at the same time there are demands for the future—things we keep trying to guess at. You know, what are the new genres of the scholars going to be in the future and how do we support them? It is tremendously difficult to do both of those at once. Neither one of them is difficult individually; doing both of them is very hard.

SPARC, at least to me as an outsider, feels like something that is largely targeted at the present crisis as opposed to the future, but I think there is an important sort of secret message inside for the future, too. This is one that the networking people learned. They built the Internet, but then it commercialized. Higher education took up a minuscule fraction of the use on the commodified Internet and as a consequence lost its ability to manage its own research agenda. That is the reason why Internet2 had to be built. Higher education was starting to envision applications it wanted to deploy on the Net, but, when talking to the commercial network surface providers, who are basically people of goodwill and are interested in these advanced applications, they would nonetheless answer, We’d like to do that, but we have a little problem. You’re about three percent of our use. The other 97 percent of our use has grown ten percent a month, and we’re having some reliability problems with the network to boot, and we’re not really inclined to destabilize it right now just to help you all out.

So that was some of the impetus towards creating Internet2. It allows, in some sense, the research and education communities to regain control over planning their own research and advanced deployment of applications.

One of the things I find intriguing about SPARC is that, if it works, we may be regaining a bit of control that will allow us, once these venues are established, to transform them considerably more quickly than some of the existing commercial products that won’t necessarily see that transformation as desirable, or at least not in the aggressive way as we might like. So I guess I agree with you. I see SPARC as something that is primarily oriented to the present, but something that may have valuable positioning for the future, as well.

But let me return to the main question that I think we want to focus on today: What is our collective response to this Internet2 opportunity? I don’t think that something like SPARC is the answer to that question. Certainly SPARC, at least as I understand it, doesn’t depend on Internet2 as a prerequisite, although of course it could make good use of it. What sorts of things are we doing or are we planning to do, or, more to the point, have we always wanted to do but couldn’t, that call for Internet2? Keep in mind that there’s something else on the table here with Internet2, as well. We are at a fairly unique point for the next year or so, where there is a reopening of the dialogue between the network designers and the application designers.

MR. KOBULNICKY (University of Connecticut): One of the difficulties in envisioning the digital library that is successful for users is trying to match one’s physical ability to scan objects and meet the distribution needs for visual information. We’re looking at the ability to move through large amounts of information that have a visual component—things we can read or images we can scan. We want to have them come along the network in a reliable way so that users can very quickly pick and choose and scan. We begin to get a feel to digital information that is a lot more akin to what we know in our present environment and which we as human beings are designed to process.

So I would argue that where we have large digital collections that we want to move to users, Internet2 might be an important vehicle for making that happen.

MR. LYNCH: I really agree with that comment; I’d frame it just a little bit more broadly. A lot of our interface design and thinking about how to deliver content for the last decade has been very much shaped by the limitations of current networks and we have, in fact, built many of our user interfaces to compensate for shortcomings in the network.

There is a tremendous opportunity to deploy some experimental prototypes that are built on the proposition that you have very fast, reliable display out there. One of the shortcomings we’ve heard again and again about the digital delivery environment is problems with scanning and browsing, and it’s certainly time to reopen the question of how to design better ways to do that. These close-to-home things are very legitimate experimental questions to explore, too.

MR. BISHOP (Northwestern University): Cliff, I was a little concerned at the beginning when you spoke about not having access to commercial services and my assumption has been that that is something we will have to address. Although it’s not a commercial service, having access to JSTOR, for example, through Internet2 would make it a great deal more practical.

We’re trying to decide—or I am—whether to put the sources up locally and deal with those associated problems or whether we should instead build a basic strategy that relies more on the network. And if we’re going to try to do that through the commercial Internet, we have some major problems.

To what extent do you think this prohibition on commercial service is something that is set, or do you think it is negotiable and something ARL should be working on?

MR. LYNCH: Let me answer that as honestly as I can. I certainly think that when we talk about non-profit kinds of content, for example, something like JSTOR, it’s a lot easier to think about getting that homed on Internet2 than on the commercial services. To the extent that you can also make the case that we’ll really learn something about getting it hung on the Internet2, as opposed to just the more convenient routine application, the case is tremendously strengthened.

I think that it would be of value for ARL to discuss with the Internet2 folks about this particular set of issues. I don’t believe the policies are by any means cast in concrete at this point and, certainly to the extent that you have compelling and experimental applications, there is a lot of flexibility.

I can tell you what is motivating a lot of the concern: The same old acceptable use policy kinds of issues that made us so crazy back in the NSFnet days. Basically, there is a real concern about the federal funding subsidizing commercial kinds of activities on the net and that’s a lot of what they’re trying to stay away from.

MR. NEAL (Johns Hopkins University): I wanted to make two quick points. The first is in response to what Harold said. It is a point I agree with in terms of the need to reach out and be visionary in our thinking about applications for Internet2. But I’m also drawn back to the immediate situation in which many of us find ourselves, where we put up reasonably progressive tools and capabilities and then we find part of our campus community saying it’s not enough, it’s so conservative, that we aren’t moving the campus along in the visionary aggressive ways. Then I have the other side of the campus asking, Where is the Lynx version? There is tremendous diversity in the community. As we look at the Internet2 applications and the visionary approaches, we want to build that on the basis of the communities that can take advantage of it and that can work with us to pull it forward, while at the same time we need to build up the campus infrastructure in other disciplines.

The second point I wanted to raise is that in the university settings we’ve recently had and are having experiences with vBNS implementations. How many universities here have vBNS or are in the process of implementing it? Maybe about a third. How many of those universities have moved forward with vBNS with the anticipation of significant library applications on that?

Well, one of the ways we sold the vBNS application at Johns Hopkins was by putting forward a few projects, which we call digital projects. And now that we have vBNS, we are having a very difficult time getting the administrators of that network capability to come back and work with us to actually activate those, because we think we need them.

The three examples we gave may respond to your question of how you are going to use Internet2. One is, like many of us, we’ve moved major resources out to remote facilities; we have a half million volumes 20 minutes away from campus. Faculty are telling us that that works okay, but they want to be able to browse those holdings. So picking up on Paul’s point, four faculty in our engineering school whose focus is robotics are now working with us on a prototype system that will allow a user to get on the online catalog and then have an item robotically withdrawn from the collection. Half the page is available online to be viewed, and we would then like to have a scanner automatically digitize that text and bring it over the campus network to the user’s office or home. When you globalize that you begin to see some really interesting applications and in the process capturing the metadata information from its source.

Two, we have on our medical campus huge amounts of visual images of different diseases and conditions. One of the things we’re testing is the ability to search—rather than by standard terms, standard taxonomy—by actually using color shading as a search strategy or tool, and then we’d like to make that available globally.

The third example is our recent conversion of 30,000 pieces of sheet music into a database that is now visually accessible on the Internet. A member of the Peabody Conservatory has written a software program that automatically converts the musical notations to digital sound, and we want to now add that capability to the database. However, we don’t have the infrastructure to deliver that locally, much less nationally and internationally.

So have we captured the attention of the people who are working on the vBNS? They are busy now linking it up into the astronomy department, the computer science department, the physics department. The things that we are advancing, although they are interesting in our community, don’t capture the attention of the people who are leading this effort. That is a real challenge.

MR. LYNCH: That’s part of the whole local infrastructure bind.

MS. KRANICH (New York University): I want to build on what Jim said. I hope that we start thinking in the bold arena, and not just about the facsimiles of what we have in print, but how to use the transmissions capability to look at video, audio, and visual, integrating those collections that are now separate, both in our libraries as well as in other cultural communities, such as the museum community with which we now have a new relationship. We need to ask, How can we be leaders in pulling together all these different formats and delivering them?

MR. LYNCH: That’s an important point. Think of how much video we saw today and one of the questions I didn’t get to ask is: Who manages that video, where is it stored, and who’s taking care of it? You know, one very inexpensive application is to start videotaping or digitally capturing video seminars and other things that are happening at your institutions. Those are the ultimate fugitive materials. They are even worse than manuscripts because they are not recorded at all except for the notes somebody takes. But they are of considerable interest around the country. Partnering with one or more departments to capture a seminar series could be a useful way in.

MS. KAUFMAN (University of Tennessee): We heard in the last panels about a number of our scholars who need to integrate and have access to huge databases. The big challenge for us is to think about that in ways we haven’t before. A selling point we used in one of the projects was the integration of our traditional and non-traditional services to reach our global-user community. One example is a virtual reference service with three or four of our institutions to serve our shared constituencies. Using those kinds of applications could be thinking slightly, not radically, admittedly, but slightly out of the box.

MR. LYNCH: How many institutions would it take to be able to have a 24-hour-a-day, 7-day-a-week realtime reference service without it being an undue hardship? It is tough for any one institution to do, but for a group....

MS. BENGTSON (University of Washington): I would like us to follow up on what our speaker Sherry Fuller said today about service provision at the point of need and integrating our services with other systems that our users, researchers, and clinicians use.

MR. LYNCH: A growing trend is that people building other applications want to essentially integrate library services as a part of them. Doing that creates a demand for you to think about the interfaces you put on your system in a very different way. Certainly, though, the bio-medical and health applications are probably the most compelling ones I’ve seen like this, but I believe there will be others as time goes on.

Another theme we heard in a number of the presentations is of geographic integration done in a way we’ve never seen it before. It takes parts of the logical set of context, the multiple institutions, and really makes them work as a coherent whole. The vast majority of the experience that we have had as a community so far has been to achieve that by centralizing it somewhere, which, as you know, is a time-honored and fairly effective strategy. I suspect, however, that that will become less effective in many of these new applications.

MR. NEAL: As we’re running low on time, I think we should turn the discussion towards what ARL’s response to this situation might be. Is there something that ARL should be doing to more dramatically influence appropriate policy issues and applications development? How do we activate that collectively rather than institutionally? How could we move forward, if we should move forward?

MS. EATON (Pennsylvania State University): A number of us had a follow-up discussion with Doug von Houweling about his comment that libraries are not at the table and not being represented in the applications area. We asked why he thought that was the case, and, as the discussion evolved, he finally agreed that it is partly the way the input has been structured. The nominations come through the campuses and they look to the project-oriented scientific projects rather than to infrastructure improvements across a national body of libraries that will then allow you to move into the applications.

Restructuring that representation is something with which ARL could help us. The question Doug asked in return was what areas we would focus on if libraries had that kind of infrastructure in Internet2. We came up with four, all of which we heard about today; I would add a fifth. I would focus on 1) a new version of e-journals, whether at the article level or beyond; 2) multimedia; 3) data, libraries, and archiving; 4) the learning center environment; and 5) new archiving technologies in general and collectively. Just as we do resource sharing, those are things we could develop and leverage in groupings.

The third thing I would like to mention about that discussion was that Doug finally understood that many of these things are funded on a project-base, yet the kinds of changes we need do not always lend themselves to project funding; perhaps ARL could help us with that in terms of policy.

MR. LYNCH: Those are all good points, and I really want to stress that first one. The way a lot of the guidance of Internet2 was populated through the participating campuses made it so that there was no obvious way for libraries as a cross-cutting interest to emerge, and I’ve had several conversations with him over the past six months along the same kind of theme.

I think he’s interested in doing something about that, and I think working a connection through ARL may be a very effective way to do so. Certainly, to the extent that CNI can also be helpful, we would welcome such a connection, as well. The funding issue is a big problem for the very reason you mentioned, that much of what’s going on now is project-oriented funding as opposed to the infrastructure-building funding that would be aimed at the overall good of the community. If anybody can take that on, ARL is probably the one, but it’s really hard to know where to start.

MS. GWINN (Smithsonian Institution): One thing we could do for each other is brainstorm about the kinds of concrete projects Jim Neal was describing; trying to think out of the box and beyond what we’re doing today is difficult when you’re back in your home territory dealing with all of those problems.

For instance, we are thinking about a music collection that might be digitized. It hadn’t occurred to me that we might actually combine it with sound—what a wonderful notion! Some more sharing and learning about those kinds of things might get us all to think more creatively about the kinds of things we might put together.

MR. LYNCH: On that theme, how many of you have had the opportunity to see, either in conjunction with Net98 in the spring CNI meeting or elsewhere, a demonstration of high-performance applications? Not too many. That’s something you really should try and find the opportunity to do. The Internet2 people put them on a couple of times a year, where they just bring in approximately a dozen high-performance applications and showcase them. Just seeing what some other disciplines have done with the technology is really eye opening.

MR. FRAZIER (University of Wisconsin): I’d like to suggest the model of the working group as a way of addressing the concept of Internet2 applications emerging from libraries. It could be under the umbrella of one of our standing committees within ARL. Another possibility is we might consider outreach to ALA and see if there is a possibility of forming a working group that would be more inclusive of other working models.

The working group model worked for us to develop the SPARC idea. It is a way of engaging in a sustained and focused discussion, and a kind of clearinghouse as well for the good ideas that are already out there. So I suggest we consider forming another working group or using the committee structure to address the issue.

MR. LYNCH: I can’t really say anything intelligent about the committee structure, but I do want to make a comment about the suggestion to link up with ALA. I think that there’s a fairly high correlation between the institutions represented in ARL and those that will be connected to Internet2. Certainly, not all of your institutions are Internet2 institutions, but I think it’s as close a correlation as you’re going to get in the library community. One of the problems you’ll face with outreach to ALA is that the people who are there and not here are not going to have access to Internet2.

As you may have gathered from some of the things Mark Luker was saying, there is a sort of queasy problem emerging about the technology transfer and how what we learn about deploying the applications in Internet2 moves past the very small group of one hundred institutions, and more broadly into the higher education and library communities.

I think they’re mostly focused on the higher education community, and one place where you might be able to work well, with NTTF in particular, is technology transfer strategies as we think about how high-end applications might move over the next five years into the library world.

MR. HEATH (Texas A&M University): I would like to address the funding issue. I spent the last week trying to nail down my subscription coverage for the next two years. The good news is I managed to do that. The bad news is that I calculated that the cost of maintaining essentially a static subscription base over a five-year period would be $2 million. I may as well have put the $2 million in a barrel and dumped it in the McKenzie River. Our dues through ARL is a five percent average increase over the past decade; they do go up. If SPARC works—if it staggers the beast—we have the opportunity to address Harold Billings’ paradigm with those dollars, to come to Internet2 as a partner with resources to explore the 21st Century. That is how the two things join: Internet2 and SPARC.

MS. FULLER (University of Washington): The other potential source of money for ARL and the individual universities is the Next Generation Internet. The amount that’s being discussed for 1999 is over $100 million, and there’s no reason why that can’t be a source of funds. In fact, this can be an almost better source than Internet2, which has such a strong technical infrastructure component to it. There’s an indication that Congress will be supportive of funding for both 1999 and the year 2000, and there should be more in 2000 than in ’99. That is something we could look to as a collaborative opportunity with ARL institutions.

MR. NEAL: Thank you all for your comments, and, Cliff, thank you for helping us think through these issues.