I could boil this column down to three sentences: Using the Web should be easy. Libraries offer complex, full-featured information services. Simplifying and integrating are hard.
We all know that libraries compete against other services on the Web when it comes to providing information. If users find it difficult to obtain information that is adequate for their needs from their library's Web site, they will go elsewhere. Google epitomizes our competition, with its simple interface and a breadth of coverage spanning more than 3.3 billion Web pages, yet able to return amazingly useful results. It seems to me that Google has mastered the art of making something incredibly complicated look simple. Online booksellers, like Amazon .com, compete head-to-head with libraries on putting books into the hands of readers. While relative to Google, Amazon is a fairly complex system, it is one that remains smooth and simple to operate.
We in libraries often have a large variety of different, but interrelated, services available for our clientele. A typical academic library will likely offer at least half a dozen different Web-based applications, each providing a specialized service. Each of these applications may come from a different vendor.
Web-based services offered by a library might include the online catalog of its automation system, an intorlibrary loan or document delivery system, an OpenURL-based reference linking environment, a metasearch interface, electronic reserves, a virtual reference service, remote patron authentication, and/or local digital library collections, as well as a plethora of subscriptions to citation databases, full-text resources, and electronic journals. Having such a variety of services and resources increases the complexity of the overall environment, and makes it harder to integrate everything into an environment that appears simple to our users.
The library Web environment isn't just a simple information-retrieval interface; increasingly, it's a place where our users receive services as well as information. Its scope includes both providing electronic information and facilitating access to traditional, physical materials. Our Web visitors expect to be able to find and view full-text electronic information as well as to renew books that they previously checked out; they expect to ask reference questions and receive prompt answers provided electronically from a librarian. Through our participation in resource-sharing systems among regional consortia and national interlibrary loan systems like OCLC, if we don't have the book a user needs, we can get it from another library practically anywhere in the world. But while libraries do a lot via the Web, most of us haven't yet achieved the simplicity of our commercial competitors.
I think that most of us would affirm that it is important to make things easy for our users, even if it means more technical work and expense behind the scenes. For as long as I have been involved in designing and implementing network-based applications for libraries, I have attempted to follow this premise. One doesn't need to have expertise with computers in order to use a well-designed system. In today's Web environment, creating a simple-to-use interface is especially important. And while it's fairly easy to create a user-friendly, single, standalone system, the challenge lies in creating a simple environment that comprises multiple interrelated systems, each produced by a different company or organization.
As librarians work toward providing new services in their Web environments, they often end up with Webbased applications produced by multiple developers. The applications may follow different user interface conventions, and may or may not be designed to integrate with the other systems the library owns. Since there exists no single vendor or product that can provide a comprehensive automation package, it's a given that most libraries will end up with Web-based applications from different sources.
Librarians today find themselves needing to be system integrators. As we acquire multiple products from different vendors, we have to try hard to make them all work together in a seamless way. That is, library Web site users should be able to navigate easily among the various services we offer without having to be conscious of the technical systems that underlie them.
Removing the Seams
While it's easy to posit the view that library Web sites should be seamless, what are some practical ways to smooth out the bumps that our users experience? Some options follow.
An increasing number of the services that libraries offer require users to authoritatively identify themselves. We call this process authentication. While some of the services need only confirm that the individual is a legitimate member of the library's clientele, others require personal authentication. Access to a licensed database, for example, can be enabled for anyone associated with the subscribing library. Viewing the books currently checked out, performing renewals, paying fines, or requesting materials all require personal authentication. Given library values related to privacy, the authentication must be secure and authoritative.
An important element of achieving seamlessness is implementing a single point of authentication-not only throughout the library's own resources, but also within the larger organizational environment. First and foremost, library patrons should not have to authenticate themselves multiple times as they use our resources. If a user signs in to view his record in the online catalog, we shouldn't make him sign in again if he requests an item through our interlibrary loan system. While it is a technical challenge to pass authentication credentials from one application to another in a secure way, it is a goal to strive for.
It's also better when the library can leverage existing authentication systerns rather than providing its own. In the academic setting, for example, students shouldn't have to have library usernames and passwords that differ from those they use for their campus e-mail system, courseware system, and the like. Libraries should work toward being consistent with the authentication schemes of their larger organizations. From a technical perspective, it is preferable to acquire systems that can operate with external authentication servers, rather than those that demand their own accounts and passwords. When it's not possible for a given application to use an external authentication service in real time, a secondary option is loading usernames and passwords in batch into the local system, and keeping them synchronized. Either way, the end user has to learn only one set of security credentials to use for all the library and other services and resources available to him throughout the organization's Web environment.
Another key element in the seamless library Web environment is providing a consistent look and feel. While the underlying software controls most aspects of the user interface of a library application, it's usually possible to customize many of the elements of presentation. At minimum, applications should allow the use of library-supplied cascading style sheets, so that, at a basic level, all the Web pages generated by each application adhere to a standard set of fonts, color schemes, link styles, and other cosmetic features. It also helps when the library can display its own graphic logo and carry the same headers, footers, and sidebars from its basic Web site over into the pages generated by its online catalog and other applications. Providing consistency among the basic pages of the library's Web site and the pages generated by each library application will go a long way toward creating a unified Web presence for the library.
Anticipating Web Services
While there are practical techniques like those mentioned above for making the library's online environment more seamless and unified, the usual approach of many library applications constrains the level of interoperability possible. At a basic level, most applications today insist on taking control of the user interface. It might be better if applications could work more behind the scenes, operating as plug-ins to other applications without introducing their own user interface components and complexities.
The emerging Web Services architecture offers greater promise for seamlessness. In a "Service Oriented Architecture," applications communicate with each other using data structured in XML over protocols such as SOAP (Simple Object Access Protocol). Without going into the technical details, suffice it to say that this approach would allow a single application-maybe the top-level Web portal-to provide a comprehensive user interface, relying on a variety of lower-level applications that work behind the scenes to provide information and services as needed. It's also unlikely that Web Services will permeate the world of library applications any time soon. Nevertheless, I would suggest that readers follow developments in this area. I believe that Web Services will eventually have a large impact on the library world, though it may take a few years.
Today, we live in an environment where it takes conscious integration efforts to make multiple library applications work together well. For the future, we can only hope that interoperability will be more intrinsic as a result of the Web Services architecture. Either way, the key goal remains to provide a simple, seamless interface for the complex set of services the libraries offer through the Web.