MyLibrary: A Copernican revolution in libraries


"We are suffering from information overload," the speaker said. "There is too much stuff to choose from. We want access to the world's knowledge, but we only want to see one particular part of it at any one particular time."

The speaker was part of a focus group at the North Carolina State University (NCSU), Raleigh, back in 1997. I was then a member of the Digital Library Initiatives Department, and we were researching what students, faculty, and staff needed from digital library services. This response about information overload was one we heard repeatedly from all segments of our users.

This was in the midst of the dot-com boom. Hot Internet real estate offered access to content such as the weather, stock quotes, syndicated news, and searchable indexes to collections of Internet resources and newsgroup postings. Additionally, these sites offered customizable interfaces to content, interfaces that allowed people to specify, for example, the types of news they wanted to read. Furthermore, these sites would "remember" you and display the same sets of information the next time you visited. These user-driven, customizable interfaces to sets of information resources, or portals, were most often called "My" services such as My Yahoo! and My Netscape.

To address the issues raised in our focus groups, the NCSU Libraries chose to create MyLibrary, an Internet-based library service. It would mimic the commercial portals in functionality but include library content: lists of new books, access to the catalog and other bibliographic indexes, electronic journals, Internet sites, circulation services, interlibrary loan services, the local newspaper, and more. Most importantly, we designed the system to provide access to our most valuable resource: the expertise of our staff. After all, if you are using My Yahoo! and you have a question, then who are you going to call? Nobody. But if you are using a library and you have a question, then you should be able to reach a librarian.

MyLibrary's underbelly

MyLibrary is a database application with a web front-end. Since database applications excel at keeping track of lists of things, we created our solution with a database application (MySQL) in order to keep track of our lists of Internet resources. First each resource was "cataloged" with its title, URL, and a simple scope note. Each resource was then "classified" using locally developed controlled vocabularies (essentially a subject classification and a format classification) as well as a few qualitative judgments assigned by librarians.

One of the controlled vocabularies, the subject classification, was critical to the success of the project. This classification was germane to the local institution and represented the disciplines supported by the university. They were subjects with which the patrons could readily identify and corresponded to the colleges and departments across campus. This classification was the glue that would ultimately be used to make relationships between patrons and resources, between resources and librarians, as well as between librarians and patrons.

Second, the names, email addresses, telephone numbers, and URL's of librarians were added. They too were classified using the same locally developed subject classification scheme. Yes, we classified librarians with subject terms, the same subject terms used to classify the Internet resources. We 'cataloged' the librarians."

Third, a set of static as well as dynamic HTML pages-a website-was created by generating reports against the database. This meant that the initial MyLibrary screen was dynamically created and presented patrons with the most up-to-date information. As the patron drilled down into the site, more of the content became static lists of resources. These were updated by the database application only when the underlying content changed.

The pages presented the content of the database in various forms: lists of content ordered by subjects, lists of content ordered alphabetically, lists of recommended resources, links to the hours, links to the virtual reference chat, links to interlibrary loan forms, etc. The pages even included lists of new items added to the catalog. This list was updated daily by extracting lists of new acquisitions from the catalog and importing them into the MyLibrary database. We only kept three months of new acquisitions data in the MyLibrary database; as new acquisitions were added the old acquisitions were deleted. In this way we were able to create a virtual new bookshelf.

Patron customization

The final piece of the puzzle was the patrons. Normally, the interface provided patrons with access to the total collection, but the access was generic because the patrons' interests were not known. But with MyLibrary, if a patron desired, they could create an account.

In addition to their name and email address patrons were asked to classify themselves by selecting one of the locally developed subject terms, such as Forestry, Life Science, or English. Once selected, the system returned a set of information resources specifically designed for that subject and thus for that patron. Moreover, the system returned the name, email address, and telephone number of the subject specialist. Lastly, since few people are interested in only one subject, the system allowed the patron to delete any resources from their page and add any others from the system or the Internet in general.

In the end we provided a means for patrons to create their own library. They had access to the total content of the system, but only the items they were most interested in were prominently displayed. For example, a student coming to the system for the first time might be presented with links to the library's catalog, an encyclopedia, a generic bibliograhic database such as EBSCO's Academic Search Elite, a comprehensive list of resources, and the virtual reference chat. When the student creates an account and identifies themselves as interested in psychology, the page's content changes to highlight PsycINFO, Web of Science, the librarian in charge of the psychology resources, lists of new acquisitions in the area of psychology, as well as the links to all resources.

Growth in libraries

By late 1998 the NCSU libraries had developed version 1.0 of MyLibrary. In 1999 the software approached version 2.0 and was being distributed as open source software. Since then many librarians have downloaded it, and to date at least a few dozen libraries across the world have implemented it as a production service. It has also become the model for many similar implementations that provide a user-driven, customizable interface to sets of library resources. The system's concept is becoming so popular that the term "MyLibrary" (coined by Keith A. Morgan , NCSU Libraries) is slowly making its way into the library vernacular.

MyLibrary's development and distribution moved to the Digital Access and Information Architecture Department of the University Libraries of Notre Dame in 2002. With this move the emphasis on customizability continues, but more work is being done on the system's ability to syndicate the system's content for other purposes and to better classify information resources. MyLibrary is still a portal application, but more effort is being spent to allow the system to work with additional applications both inside and outside library settings.

Syndicating and repurposing content

Libraries are always a part of a larger organization. Public libraries are a part of municipalities. School libraries are a part of K-12 institutions. Special libraries serve governments, hospitals, or companies. Academic libraries are a part of colleges and universities. Because of these relationships librarians need to seamlessly integrate their content with the content of the larger organization.

With this in mind, we are exploring how MyLibrary can export and import its content to and from external sources. For example, nobody likes maintaining data in more than one place. Since much (not all) of what might appear on a library's website (and therefore MyLibrary) also exists in a library's catalog, efforts are being made into learn how MARC records can be exported from an integrated library system and imported into MyLibrary. For that matter, maybe MyLibrary could create MARC records intended for the library catalog.

MyLibrary is being used to syndicate content in much the same way news stories are syndicated across the Associated Press news wire. To this end, MyLibrary will improve its ability to output various formats of XML including RSS (Rich Site Summary) files, RDF (Resource Discovery Framework) files, and OAI-PMH (Open Archives Initiative-Protocol for Metadata Harvesting) streams.

These formats can then be picked up by external sources and distributed for other purposes. The RSS files can be read by the increasingly popular RSS news readers. The RDF files can help support the Semantic Web. The OAI-PMH streams can be used to share metadata with OAI-PMH service providers for the purposes of facilitating things like union catalogs, alerting services, or global searches.

In the University Libraries of Notre Dame, MyLibrary has been demonstrated as another source of syndicated content for the campus-wide information systems known as uPortal and SCT Luminus. Both of these systems (portals) aggregate content from across an institution of higher education. They are intended to be "one-stop shopping" systems providing direct access to things like a student's grades, schedule, financial status, course reserves, email, and associations with various groups. Here, MyLibrary is a means to insert patron-and library-specific-information into these portals, giving the student or faculty member a more holistic view of their information universe and reducing the set of information silos they need to manage.

User-centered classification

Future implementations of MyLibrary will allow the classification of information resources to be more flexible and more user-centered. Presently the system allows for only two different classification systems: subject and format. The subject and format classifications are 100 percent defined by librarians, the subject classification is unlimited in the number of terms that can be used. The format classification is limited to five terms. Most MyLibrary implementations I've seen use something like "Library Links," "Community Links," "Reference Shelf," "Bibliographic Databases," and "Electronic Journals" as their format terms.

Based on usage, it's clear that other classification schemes-controlled vocabularies-would be beneficial, such as audience, ease-of-use, time, education level, place, authority, popularity, and more. Consequently work is being done to allow librarians to create any number of classification schemes and fill each scheme with institution-specific terms. Thus, each information resource will have the ability to be described with one or more terms/facets particular to the patrons' local environment.

We all know that some Internet resources are easier to use than others. Google is simple. Some databases are very complex. Some books are written for one audience versus another. There are children's books, books for the average reader, and dissertation-level materials. The same holds true for Internet resources. Some information resources help solve one type of problem and others help solve different problems. Compare and contrast the functions of a dictionary, encyclopedia, directory, index, and catalog to see what I mean. Likewise, the resources available on our library websites need to reflect the different ways patrons approach information gathering and use. They need to be organized from the patrons' point of view, not the librarians'.

Becoming patron-centric

Patrons may not be able to articulate their information needs, but they can describe themselves. They know whether or not they are a child or an adult, a student or faculty member. They know what they are experts in and where their knowledge is lacking and whether they seek a simple answer or a more authoritative, comprehensive one. Patrons know how much time they have already spent on their problem, and they know how much time they are willing to spend. Because our users can describe themselves, we need to organize our resources to coincide with their descriptions, not the other way around.

The terms and facets describing information resources should be user-centered and not derived from centralized authority lists outside the patrons' immediate environment. Instead of describing information resources based on foreign, remote, objective, systems like Dewey or Library of Congress Subject Headings (LCSH), MyLibrary will allow librarians to describe information resources based on concepts local to the patrons' environment. This will make more explicit the information they need to know. Dewey and LCSH are controlled vocabularies intended to meet everybody's needs, and by trying to be everything to everybody, they serve few people very well.

One size does not fit all, and in order to provide better access by individuals to our services and collections we must present these services and collections in terms that individuals can literally relate to. These terms need to come from our patrons' work and education environments, their time restraints, their experience and expertise, their personal characteristics. I don't advocate getting rid of Dewey nor LCSH. Instead I advocate supplementing the descriptions of our materials with additional subjective, qualitative, institution-specific vocabularies such as audience, authoritativeness, ease-of-use, complexity, and timeliness.

You might ask, "How do you develop these user-specific terms? How will these terms be collected, and how would they be used?" The answer is not difficult as long as you keep two things in mind. First, MyLibrary is not intended to be a comprehensive inventory list of all things accessible from a library. Libraries already have a tool for that purpose, and it is called the catalog. Instead, MyLibrary is intended to much like the pathfinders developed at MIT in 1965. MyLibrary is intended to be a system for managing and disseminating lists of highly regarded and most useful information resources. The content of MyLibrary is intended to be the same sort of content of pathfinders, and remember, pathfinders were, by definition, short. MyLibrary is intended to be a proactive finding aid, not a catalog. Since the items in MyLibrary are intended to be selective and not comprehensive, the numbers of items that require user-specific terms are relatively small. The second thing to keep in is that you must know patron, and you must be able to describe their information requirements based on their personal characteristics. Ask yourself, "What qualities describe my patron population?" and you will be well on your way to creating a user-specific list of terms.

For example, suppose you work in a public library you have a set of patrons. To a greater or lesser degree, these patrons have qualities such as but not limited to age, educational background, physical capabilities and limitations, the desire for a particular format, technical abilities, information needs, time restraints, and subject expertise. Each of these qualities (think of them as facets), can be exemplified through the creation of a list of terms. The easiest one for our profession to get our minds around is subject expertise. Subjects can be easily enumerated: astronomy, philosophy, music, literature, history, economics, etc. Information needs usually include things like the need to know the definition of a word (dictionary), the need to acquire a short outline of a topic (encyclopedia), the need to identify a list of current articles on a particular topic (bibliographic index), the need to know the name, address, and telephone number of somebody or something (directory). In other words, information needs can often be met my using specialized tools designed for that purpose and these tools can become terms in an information needs facet. Besides dictionary, encyclopedia, index, and directory, other terms come to mind such as manual, thesaurus, catalog, concordance, atlases, or almanacs. Particular formats can be a facet, and terms exemplifying formats include things like books, scholarly and magazine articles, images, sounds, movies, data sets, microfiche, manuscripts, or maps.

A more difficult facet to enumerate is educational background. Some information resources are designed for the scholar and require a higher degree of education to get the most out of them. Therefore terms used to elaborate upon education background and describe information resources might be grades K - 6, 7 - 12, college student, and professor. People know what grade they are in, and by classifying resources by grade, the resources most useful individuals will more likely bubble up to the top. Similarly, some resources are easy to use and others require more time and energy. Terms to exemplify time constraints might be novice user, intermediate user, advanced user, or expert user. People who define themselves as person with little or no experience might be steered towards the novice or intermediate resources before the more difficult-to-use or time-intensive resources are suggested.

MyLibrary assists the librarian in the creation of these facets and terms. MyLibrary allows the librarian to create any number of facets (audience, subjects, types, media, time restraints, etc.). It also allows each of these facets to be elaborated upon (freshman, sophomore, junior, senior, astronomy, philosophy, literature, mathematics, dictionaries, encyclopedias, directories, indexes, catalogs, books, journals, maps, data sets, images, sounds, 1-10 minutes, 10-30 minutes, 30-90 minutes, greater than 90 minutes.). Once enumerated, each resource in MyLibrary is classified with one or more of the facet/term combinations. Finally, MyLibrary learns about the patron. It learns how they exemplify a part of the audience, what their subject expertise is, what kind of information they desire, and how much time they are willing to spend to acquire it. Based on this information MyLibrary identifies information resources pertinent to an individual's needs and desires.

Meeting expectations

The Internet and our society's move towards a service economy has changed the way people use, gather, and disseminate data, information, and knowledge. Now, more than ever, libraries must compete for people's attention. People's expectations have changed, and people are bringing these expectations to the library. They say, "Why can't you make it easier? Google does." As I listen to students and faculty here on the Notre Dame campus, I increasingly hear , "I want a portal. I want a thing that will allow me to create a customizable view of the resources I am interested in. I want a service that is like where things will be recommended to me based on who I am and what I've used. Library, can't you do that? " With MyLibrary, such things are entirely possible.

Whether we like it or not, our society is increasingly driven by convenience. Even though we understand that some things are the "right" things to do, like exercising and eating correctly, we don't do them. That is often the case with library use. People know the information they desire can be found in libraries, but our systems are too difficult to use. They are too difficult to relate to. When people think of libraries they think of the Dewey Decimal system.

By designing our library websites from the patrons' point of view, and by trying to meet the patrons' Internet expectations, the library profession will be more likely to meet its patrons' education and information needs and desires. Such a process does not belittle the profession nor deny patrons the ability to become more information literate and self-sufficient. It just addresses the issues from a different angle.

Is MyLibrary a "Copernican revolution" for describing information resources? No, probably not. On the other hand, by creating local classification schemes rooted in the experience of the patrons and then describing information resources using these schemes, MyLibrary does turn classification a bit on its head. Instead of the classification process revolving around objective, one-size-fits-all terms, MyLibrary might allow classification to revolve around the patron and their personal knowledge and characteristics. Only time will tell.

Creator: Eric Lease Morgan <>
Source: This is a pre-edited version of an article published as "Putting the 'My' into MyLibrary, netConnect (Suppliment to Library Journal), Fall 2003, pgs. 24-26.
Date created: 2003-09-08
Date updated: 2004-12-12
Subject(s): articles; MyLibrary;