2. http://www.niso.org/committees/MS_initiative.html “Meta search services rely on a variety of approaches to search and retrieval including open standards (such as NISO's Z39.50), proprietary API's, and screen scraping. However, the absence of widely supported standards, best practices, and tools makes the meta search environment less efficient for the system provider, the content provider, and ultimately the end-user” Having said that, there is hope for Meta search. Among those interviewed for the LC report, there was some hope and many fears about Meta search as a technology, but no consensus. Comments ranged from “meta search is a fatally flawed technology” to “Meta search may not be the right solution but it is addressing the right problem” to “Meta search has enough promise that we should go forward with it.” Among the many interviewees who talked about Meta search, there was agreement that the NISO Meta Search Initiative is critically important to the future of this technology. The quote on this slide is from the Initiative’s Web site. The problems with Meta search are pretty well documented. Besides the absence of shared standards, which was interviewees’ most frequent complaint about meta search, they cited problems with the time commitment required for local and vendor work with meta search engines and to keep connectors working, the absence of needed relevance ranking in search results, and the nascent state of meta search technology. Google Scholar: Forget Meta search? Some writers, like Marshall Breeding, are beginning to point to Google Scholar as an example of a better approach (i.e., searching based on a centralized index). Once the information seeker finds a book or article of interest in Google Scholar, they can use reference linking to connect to the content offered by their library, which is what the slide here illustrates. My own sense is that Google Scholar, which I think is still in beta, is still some distance from a sufficient supply of scholarly content to be a real substitute for Meta search in libraries. It is also too hard right now to set up links back to one’s own library holdings—even if the library has completed its “deep linking” work, the information seeker not working within the IP range of his university has to know to set preferences, and then how to do it. I myself am thinking that Meta search will need to be around for a few more years. Reference Linking Users expect fully linked information environment Partnerships between content providers, database producers, and library system vendors, utilities … Now back to reference linking. Limitations of Reference Linking Incomplete or inaccurate metadata from source; can’t match knowledge base Knowledge base is incorrect or out of date Metadata alright but doesn’t match target Varied application of citation standards; non-use of citation standards Library has full text for journal but not the volume/issue the user wants Full text availability lags behind citation availability And on and on As for reference linking, Open URLs don’t work sometimes, and links that should be made between sources and targets are not always successfully made. But there are many possible reasons why links don’t succeed; some are listed here. The Portal Dream, Version 1: A Unifying System Model Other LibrariesCatalogsLocal Library CatalogDigitalCollectionsLicensedDatabasesOther(e.g.,DSpace)Many diverse, separate interfacesFederated searching (Meta search)Authentication layerUnified Web Interface (“Google-like”) This model illustrates a library that provides access to a rich but overwhelming array of resources. These resources might be described in the types of databases you see here. They all have separate, different user interfaces. Library users are often on their own to be aware of what online and print collections are available to them, what they contain, and how to find and navigate their many interfaces. We need to build new library systems that help users find what they need quickly, without having to sort through masses of materials and online data stored and organized in multiple places, in multiple ways. The dream of the next generation library system—sometimes called a portal--is one with an integrating layer for all of these resources. In recent years, many have thought that met search, or maybe OAI harvesting, or both, will provide the integrating tools to realize the dream. Six years ago the Cornell library became a development partner with Endeavor Information Systems to build EN Compass. This was the dream we had for a unifying system model. The underlying assumption was that we would want to integrate everything in one big, diverse, but still local information system, whose home would be our library Web pages. We have now gained some experience with how such a system might manage and integrate the diversity of resources to which libraries provide access. Look from a distance! While we were pursuing this dream with Endeavor, which by the way is still a good dream, more and more members of the Cornell community began starting their searches not on our library Web pages, but on the open Web, or on course Web pages, or within repositories like the physics at Xiv that were not managed by the library. If one steps way back and looks at the nation’s (or world’s) libraries’ separate, independent attempts to integrate information resources for a local community of users, the picture that emerges is like the nebula here. Library collections of all kinds—print an digital—and a wide variety of scholarly information resources are isolated in terms of how they relate to one another, who is responsible for them, their delivery platforms, and how standards are applied. Yet, with a bit of adjustment to our dreams, this nebula might become a factory of stars and planetary systems, yielding immensely favorable results for scholarly information seekers. Outward Integration Integration should be outward rather than inward, with libraries seeking to use their components in new ways” --Interviewee for LC report on future of the catalog A galvanizing comment for me, while doing the research for the LC report, was this one. As project manager for the EN Compass project at Cornell, then team leader for another Cornell project to prepare requirements for an integrated framework for the library’s 50-odd digital collections, I had been focused on the goal of inward integration—that is, integration of discovery on Cornell library Web pages. This comment crystallized the insight that had been growing in me that I had it at least partially backwards. Integration should be outward—in the direction of the open Web. In other words, instead of assuming that users would come to our pages, we should assume that users will be searching on the open Web, using mostly search engines, and our job was thus to make our data visible to them there, then pull them in to fulfill their needs through our local collections and delivery systems. Longer Term Vision Switch users from where they find things to library-managed collections of all kinds Local catalog one link in a chain of services, one repository managed by the library More coherent and comprehensive scholarly information systems, perhaps by discipline Infrastructure to permit global discovery and delivery of information among open, loosely-coupled systems Critical mass of digitized publications and special collections online Many starting points on the Web leading to many types of scholarly information objects In this way, my thinking shifted to another kind of dream for a unifying system, one that leverages both the strength and power of popular search engines with the wonderful assets that libraries and scholars have to offer. Here is a first attempt to articulate the components of the new dream of a unifying system, in which libraries and Integrate Library System vendors play roles, but not the only roles. Find It on Google,* Get It from that Library Open World Cat, worldcat.org Google Scholar, Book Search Google Library Project Million Book Project Microsoft Live Search Books Open Content Alliance Amazon *The word
Google
was first used in the 1927 Little Rascals silent film
Dog Heaven
, used to refer to a having a drink of water. en.wikipedia.org/wiki/Google (verb) We are seeing a lot of tinkering with the pieces of a new vision for a unifying system model. These are some of the projects that dome to my mind as to signal the approach of what is to come. These particular projects are of great interest because they involve the kinds of assets that libraries and A & I services have generally looked after, books and the serials literature. Live Search Books Cornell University Library Digital Collections Amazon/Book Surge Acquisition “The acquisition will allow Amazon to profitably market hard-to-find books which can now be produced by Book Surge in quantities as low as one.”—press release Intermediate Vision Shared OPACs: begin to aggregate discovery function for books, serials, and their e-counterparts Meta search for e-journal articles Reference linking ubiquitous Draw on the local catalog’s strongest suit: support for inventory control and delivery Larger scale collaboration on collection development/resource sharing, storage, preservation But for now, the new dream is just a dream. I believe we can get there, but it will take time, and there will be many course adjustments along the way. The shift will have many intermediate stages, as discovery begins to happen more in popular search engines or services like Google Scholar, and delivery more the domain of libraries, online bookstores, and other suppliers. The OPAC interface is more likely to be part of a shared catalog of some kind, with the local catalog and Integrate Library System serving as “last mile technology” to carry signals from and to the shared OPAC and provide infrastructure at the “neighborhood” level to complete the discovery to delivery chain. Along these lines I think we should be looking for Integrate Library Systems that are less monolithic and more open, modular, and more compatible with other systems. As these trends gain momentum, there may be more compelling reasons to share the costs of building, storing, preserving, and delivering collections—traditional, electronic, and digital—to users. Intermediate Vision, 2 Greater use of Web services to link in and out, tie applications together Start to build bigger scholarly information environments—with libraries playing a role—to aggregate more of the expanding universe of scholarly digital assets Metadata and outreach skills = strategic assets The libraries will start paying more attention to the research and learning objects that are popping up all over campus and that we are calling “digital assets.” Students and scholars are creating these assets, but generally libraries are not involved in supporting them. Some of our libraries have DS pace repositories, and some of the faculty assets are stored there, but not many. As the trend toward bigger and more heterogeneous scholarly information environments takes hold, the library will have at least two strategic assets to offer—experience with effectively organizing and preserving information on behalf of others, and knowledge of the key resources of a discipline. Intermediate Vision, 3 Beginning of the era of special collections Aggregate discovery of digital collections More emphasis on visual resources More collaboration with faculty on digital assets Rise of best practices for digital asset management Digital collection delivery platforms will continue to proliferate If we think back to the nebula again, we have the opportunity to make special collections into a lovely set of stars and planets. These collections have been hidden away, but circumstances could be such that these unique special collections will take on more importance, prestige and weight for libraries. Should this part of the dream come true, we will need to manage—or have someone manage for us--multiple systems delivering a wide variety of objects from special collections—images, text, sound and other media. Digital Collections Ralph, Julian Canada’s El Dorado Harper’s, Jan.1891. Making of America Collection let me wind up with some examples of where I think we are, and where we might go, with digital special collections, and the kinds of linked systems and platforms we will want to make them visible to users and to manage them. I explored the Cornell library’s Making of America collection for information about Canada. I found this wonderful image of British Columbia imbedded in an 1891 issue of Harper’s magazine. You can also find and get that image using Google. It is the fourth link here. It has been our experience at Cornell that users are increasingly finding objects from our digital collections on Google first, and then they are connected to a page from a finding aid or other collection, without much context for navigating the riches they have stumbled upon. Good Advice for Digital Librarians At this stage, no new effort should be undertaken without a sense of how it will be merged with other existing collections and where the resources for long-term maintenance will come from. —A CUL digital projects librarian We have found at Cornell that we are better at building digital collections than we are at connecting them to other related collections or at taking care of what we have over time. At present, our 50-odd collections represent a mix of a few comprehensive collections (like the Core Historical Literature of Agriculture) paired with collections built through one-time funding opportunities. These smaller collections haven’t got much force of attraction pull on their own Aquifer The Digital Library Federation’s Aquifer and projects like it offer best practices and tools that could someday facilitate drawing myriad smaller collections from the nebula into planetary systems that will serve information seekers and scholars better. It could be said that Aquifer is an initiative promoting the kind of outward integration that I have been talking about. Bridging Digital Islands The next generation of Integrate Library Systems will need to support the next generation of students, researchers, teachers and scholars at our universities and colleges. Satisfying their needs will require modular Integrate Library Systems that can be put together like legos. Standards for connectivity and linking, like reference linking and Web services, will be extremely important in making these loosely coupled systems interoperate. Libraries and their information systems will find ways to not only co-exist with the Ama Zoogles of the world, but to take advantage of them to expose their rich collections more effectively and to a broader audience. It is too soon to tell what the long-term role of Meta search will be. For now it is, on balance, a useful tool for leveraging our heavy investments in licensed e-content and making these resources easier to use. CONCLUSION We will need Integrate Library Systems, or at least a collection of interacting modules, that can integrate access to a greater variety of information objects and digital assets. We are entering an era of special collections and that by collaborating more we have the opportunity to make them much more visible to a worldwide audience. Integration should be outward rather than inward, with libraries seeking to use their collections in new ways.