Summary Notes, Discussion Sessions—Discovery to Delivery Symposium, March 2007
Discussion session 1: Bridging the library community with the rest of the Web community
Catalyst: Cyril Oberlander, University of Virginia
Communities: How do we balance niche audiences against the general discovery environment, and build community partnerships? How do we collaborate to create a framework for niche or specialized audiences? Niche environments will evolve organically; we need to put tools directly in the community's hands so that they can help build the environment in which they do their work.
Collaboration: How do we move from making laundry lists of desired services to actually delivering them? Can we divide labor—which bits do we need to build for ourselves, which bits can be co-opted from other environments, and who is best suited to build the solutions we need? We need to identify the problems that no one else is going to solve for us and move into that space. What things are unique to reseachers/research libraries that no one else cares about? What do our faculty, our students want that only we can do?
How can we limit the proliferation of maverick enterprises? Can libraries participate in "crowd-sourcing" approaches to building common solutions?
Can we recognize and honor the importance of failure, letting go of experiments when it's time? Libraries need to become agile—adopting, adapting, assessing. How do we get better at doing this rapidly? We need to make go/no-go decisions in weeks, not years. This demands a cultural shift—not just hiring more programmers but looking for pragmatic solutions that are already at hand. We need to decompose the large problems we face, such as online discovery, into shared challenges with relatively lightweight solutions, for example, effective disclosure of one's collections to search engines and information aggregators.
Which institutions are developing solutions and how can we learn from them? Can RLG Programs be a place for this, for example, as a repository of specialized tools we can look at? A place we can to turn for the five or six major trends and what we should collectively be doing about them? We are still trying to solve problems from yesterday. RLG Programs can help us tackle tomorrow's problems today.
Assessment: It is important to develop metrics to measure value and impact and need, and especially important to put assessment before development. Assessment lets you save successful bits, the parts that work, from what otherwise fails. RLG Programs should surface and synthesize the good work done at different institutions. RLG Programs is better positioned to do neutral assessment with evidence-based tools rather than assessment based on what we already know or have suspicions about. RLG Programs could identify and publicize the most successful bits from a variety of projects.
Branding: When our services are successfully embedded in the larger community, the library becomes invisible. How can we reassert the library's special role as a source of authoritative information?
Our environment: We have to remember that businesses and models that we look to as examples (Netflix, for example) are themselves already changing, and that the cycle of change is relentless. Network solutions are very quickly overtaken by the next wave of solutions, always pushing solutions out to the next big problems, what needs to be done next. Solutions at this level will be evolving raidly. How can we manage the downstream changes that will result at the local level?
Discussion session 2: Delivering to users wherever they are
Catalyst: Genevieve Clavel, Swiss National Library
Multilingual issues: How do we address scholarship, research, and learning in multiple languages? Some mapping can be automated using dictionaries, but most of this work can only be done manually—and it's an example of something that's best done collaboratively. Collaboration could mean capturing indexes, bringing them into a context where more people can use and access them, a pool of data that people can contribute their own terms to. We could build on indexing done elsewhere, outside of the library community. There's never a 100% solution and we have to let go of our perfectionist tendencies.
Need for registries: We need a service registry connected to institutions to register and act on cooperative borrowing agreements. Can we do unmediated requesting of e-journal articles? A catalog shouldn't just show that a digital copy exists; the information needs to be actionable and the digital copy needs to be retrievable.
Users don't stay in the same place, they don't stay the same people and don't have the same roles. Many services at the British Library, for example, serve the commercial sector. When students graduate and join the workforce, they still need materials that are licensed by their campuses. Services that provide materials for the workforce should be provided by national libraries and public libraries in combination with universities. The Perceptions of Libraries report points to this issue: students stop using library materials when they no longer have access. Registries could link people to the resources libraries pay for on their behalf.
Enriching catalog data and improving browse-before-borrowing scenarios: Are we deliverying the right things? The catalog needs as much information as possible—social tagging, reviews, tables of contents, cover art, indexes—when access by a larger audience is not mediated by a reference librarian. The more detailed evaluative content helps prevent disappointment and shipping books all over the state, country, and world for no good reason. Researchers know more about the articles they request than about books because of abstracting and indexing services. How can we have that for books?
How do we use all the bits of information scattered throughout the Web, many of them uncovered as part of different projects, and bring them to the network level where we can all make use of them? There are economic issues for data under copyright. What is the value of centralizing information about copyright? Who pays?
Consortial agreements moving to shared print: A Pacific Northwest study of large, small, and tiny institutions showed that the University of Washington, with the biggest collection, was also the biggest borrower. Many times the University of Washington borrowed copies of what it owned from its Orbis Cascade Alliance partners because its own copy was in use—it was faster to borrow another copy than wait for its own copy to be returned. If an item is not held by any Alliance partner, the library buys it, an aspect of collective collection development. In Minnesota, a state program has a delivery warehouse with its own collections budget. Materials are shipped directly from the warehouse, and are not necessarily returned to the library of origin.
Using born-digital material: We need, increasingly, to think about not only the content but the functionality that needs to be included in a digital delivery environment. Persistent, permanent links are very important for electronic access. In a physical world you deliver content to the users. In an electronic world you deliver the user to the content. The discovery environment needs to include information about what users can expect to do when they get to various versions. What expectations do users/researchers/scholars/students have? Can RLG Programs synthesize the research conducted so far? Studies from Minnesota, Washington, Syracuse, and the University of Virginia are relevant.
Issues that could develop in the future: Delivery is local if we think of things as locally owned. What happens if we think about delivery being on the network also? What happens if we share, and then delivery is a network-level activity also? Delivery issues are complicated by how much we will own and how much we will license. E-books, for example, are licensed rather than owned, which limits how much we can share.
We haven't had to worry about our collections being used too much. We could move to a situation where cooperative borrowing works so well that books are less available for a local audience. Then you start to worry about appropriate use and how more effective browse/disovery will be important in this. We'll have digital copies of materials that are not out of copyright, but are more discoverable because they are digitized, so this may increase demande on physical copies that are not currently used very much.
Who is doing work to make sure that reseachers can make interlibrary loan requests wherever they are, even from Google? Users need to be able to track the status of their requests and the expected delivery date.
Other discussion: Print on demand, distance learning, and ILL. Getting away from packaging up 40-pound boxes and mailing them to other institutions is "a tough nut"; can RLG Programs gather data on institutions that are mailing to patrons' homes, and how well it's working?
Discussion session 3: Digitization & impact on use & delivery
Catalyst: John Wilkin, University of Michigan
Rights management: Public domain issues are binary, however one can't set aside issues of fair use, which are not binary but case by case and situational. A central system needs to distribute responsibility to responsible agents who can take actions on a local level. We need a rights management framework. We can say something about copyright, something about authentication/authorization. We need to say something about the purpose—a download versus putting something into a course management system, for example. Individuals have different roles and rights in different contexts. Many projects (Google book scanning, for example) will make material more discoverable. How can we reconcile greater discoverability with the stop signs we put out to our users?
Policy management: Rights can be attached to a person, institution, or government. We need a centralized place, Shibboleth-like, to register local policies. We need mechanism that adapts to the dynamic nature of rights—when a work comes into the public domain, differences in rights depending on geographic location, and rights determined by an individual's affiliations.
Where there is no fair use allowed, someone has to pay. If this is not the library, present the available options to the user—for example, a link to a bookseller. At what point do libraries present payments options to the user?
Purpose of use: Purpose is just as important as who you are. A user's purpose may even change during the discovery process. We need rights resolvers, like we have link resolvers. We need to expose in the object the elements that help determine what rights are. Updating the resolver, rather than the object, would simplify management.
Discovery issues: There's no easy way to find digitized material currently available. Bill Carney from OCLC is working with Google, Microsoft, the Open Content Alliance, and participating libraries on e-content synchronization. Data about the works that Google has digitized from the University of Michigan will start appearing in WorldCat in June, which will demonstrate the value of incorporating information about digitized works within the aggregate of printed works. Rights information at the network level would support multiple delivery or presentation environments.
Whither the catalog: At the work level, OCLC will have significant representation of digitized content. If rights, authorization, and proposed use are all combined at a network level, why not move all discovery to the network level? Keep only inventory control at the local level. Forget about investing more effort into the local catalog.
Fee or free: How will we make digital materials available freely in the educational arena, but for charge for industries? How do we turn the tables on ourselves? What kind of citizens will we be as owners and consumers? We can make it available free, but under a Creative Commons license that inhibits use iin a for-charge context. Pressure not to charge comes from going by the "old-school" rules: make it available to people who need it. But there is a lot of pressure to charge. There is a disconnect between what we believe as librarians and where we need to move.
What is the emerging business model? Look at what's happening with university presses, which started out as a subsidized way to get scholarly publications out. These changed to profit centers on campus, and this has killed them by putting them into direct competition with publishing companies.
Making materials freely available to all may go beyond an institution's mission. It's not part of UCLA's mission to make things freely available to movie studios. Right now UCLA is working on recon rights. How do we get faculty and other authors on board with revamping rights awareness, which will help get rid of the "Disney effect" on our lives, allowing users more access to the results of research while recognizing intellectual property rights?
Obtaining permissions: Within the UK, the legal deposit system did not cover the digital world until 2003. Selectively harvesting and archiving Web sites only by permission requires enormous effort and is untenable. The British Library reported that only 23% of the sites contacted responded, and hopes that new legislation will help expand the sites covered.
The MIT Open Courseware project wanted to include content from published books, journal articles, etc. An effort to clear permissions had a response rate of less than 20%, and most requests were denied, so the effort was abandoned. Now they are recreating content so they can put it online with a Creative Commons license.
An alternate approach would be to accept silence as consent, with good take-down policies if there are objections. Not all copyright violations are created equal. The impact on the value of a work determines the severity of the copyright violation. What is considered "due diligence"? Copyright issues are also affected by the changing contexts of the users discovery environments. For example, consider a possible scenario where a user could pay 99 cents to use something, like for music now. Vendors will make content available conveniently, with one click. How many of our users will go through many steps to get this same content for free? Or less authoritative content? If materials are available very cheaply and very easily, how does this affect our rights under section 108?
Other discussion: Museum rights and reproduction issues. Making images freely available for scholars; charging for other uses.
Library Service Frameworks, assessing needs: There are expectations for not just the abstract conceptual framework but definitions that will lead to business requirements. RLG Programs partners can look to specific outcomes under our New Modes of Research, Teaching, and Learning theme. We are inventorying services and utilities that will support personal collections, understanding researcher needs and expectations, determining which ones out there are most suitable.
Rather than asking people what they want, put up prototypes and see what sticks. People know what they want when they see it. Run more experiments and assess their success.
Rapid prototyping is useful. Can we do more? Can RLG and OCLC help? Ask faculty, etc. about barriers and what the library can do to remove them. There's a hierarchy of need: if the fundamentals aren't met, you will never get to the new and interesting, but will keep covering the same ground with your constitutency.
Network-level thinking: What things can be worked up to the network level? It's important to clarify this because it helps institutions make decisions about where to focus, to see where they can free local resources for doing collaborative work on other projects.
Whither the catalog: We need a real sense of proportional investment in local discovery system improvements. The next-generation user interface is much more than just local discovery and needs links to the library's management tools. There's a real tension between tweaking the OPAC and focusing efforts elsewhere. What are the tradeoffs between using a common platform and customizing for niche audiences? For example, a science library would want science materials to rank higher in search results; researchers may want to focus on unique materials rather than the most commonly held. Could machine services provide options for individual preferences? The more discovery tools there are, the less likely users will refer to any of them.
A tuning or algorithm issue? The question about customization is core. Too much customization is a bad thing. This may be why our users have issues with the OPAC. Even as librarians we have issues when we go to to other OPACs. Perhaps customization should be done at the user level, not by the library on behalf of the user. There are two ends of a tension that we need to grapple with: tweak to meet needs versus tweaking because you can. In very many cases, we are not just tweaking screens, but actually making things work at a very basic level. WorldCat Local is a user interface. The traditional library catalog was designed for inventory control. Needs and focus of different users are different. Refocus from tweaking the user interface to building service packs for different users.
Return of branding: Consider branding versus marketing. Branding gets annoying. Our users aren't bothered by the fact that we aren't branding ourselves. We shouldn't get in their way. Can we do this without getting in their way and be creative about it? What about presenting a "library brand," rather than separate institution brands? In Google and MySpace, is putting in a brand really that annoying? There are issues of authenticity, preference, and validation of materials that come with a library brand.
RLG Programs: How do partners know what RLG Programs is working on? What are the five things they can think RLG Programs is doing? We don't want to replicate, but don't want things to get dropped. RLG Programs could facilitate collaboration among partners, provide opportunities for networking, and bring together local technological fixes and share within the partnership. RLG Programs should know what's going on and synthesize it.
Other discussion: Network level for preservation. JSTOR's experience with print archives, and willingness to share information. The lack of a bibliographic framework for archived Web content.