Searching for the Library: University Home Page Design and Missing Links

Roberta J. Astroff

Genre analysis is used to explain the placement of links to the library on more than one hundred college and university home pages. Despite the lack of established standards, university home page design falls into common patterns, indicating genre development. However, a number of university home pages do not provide a direct link to the library Web pages and thus disrupt user expectations. On those sites, the Web designers provide other access to the library Web pages either through redundancy or by classifying the library with other services. Omitting an active link to the library does not serve design principles, users, or universities well.


Almost every college and university in the United States has an official Web site. These sites appear to serve three functions: digital college catalog, public relations brochure, and access point to university online services. However, the home pages of a number of colleges and universities do not provide an active or visible link to the university libraries. Using qualitative methodologies, particularly genre analysis (usually applied to other contemporary media), this article analyzes the home page designs of the 109 U.S. and Canadian colleges and universities that are members of the Association of Research Libraries (ARL) in an attempt to understand the placement of links to the library. 1 The analysis focuses on the Web page only. It attempts to identify formal characteristics of the text itself. It does not compare the nature of university home pages created by public information offices to library home pages. It does not analyze user perceptions of Web pages or the intentions of university officials and Web page designers. It analyzes what appears to be the first gateway Web page of a university in an attempt to learn something about university home pages that do not provide an immediately visible link to the library.

Media, Genre, and Analysis

A number of researchers have begun to use the concept of genre in their analyses of digital documents and the Internet. 2 While genre is perhaps most commonly understood as a synonym for category, contemporary theories of print, broadcast, and electronic media define genre as semiological frameworks within which both the producers and users of media texts operate. That is, a genre provides a shared code, a set of expectations about the resulting media product. 3 McQuail says that ". . . genre may be considered as a practical advice for helping any mass medium to produce consistently and efficiently and to relate its production to the expectations of its customers." 4

Television audiences, for example, have learned and now expect that situation comedies will end happily, with all problems resolved at the end of the half hour. At the same time, the formula provided by the genre facilitates production, since the writers, directors, and producers are not producing a new form from scratch, impossible to do on the weekly basis commercial television demands. As it has been developed within the television industry, genre offers guidelines about almost everything from the duration of each scene to the number and placement of cameras. Attempts to change generic expectations, such as adding music to news shows or killing off sitcom characters, tend to produce controversy and discomfort among audience members. 5 Perhaps more importantly to those of us concerned about Web page design, disrupting audience expectations of a particular genre tends to frustrate that audience into abandoning the show.

But what of successful new genres? The lines between genres can blur when producers experiment, producing such "recombinant genres" as prime-time soaps, entertainment news, and TV news magazines. 6 In addition, new technologies have provided the institutional impetus for new or modified genres. Printing technology, for example, modified our expectations about the form that a book takes by allowing for the development of the table of contents and by producing identical copies of a book. 7 The development of the publishing industry and its regulation provided other standards, such as a title page, a copyright page, and publishers' information.

Dillon and Gushrowski provide a superb summary of the application of genre theory to the Web:

One of the important determinants of acceptance and use of any document type is the role it plays in supporting a discourse community. Many familiar document types have evolved over decades or even centuries of use to give rise to the highly conventional forms that are instantly recognized as being of a type or genre. Detective stories, scientific articles, newspapers, catalogs, etc. are all forms of document that have identifiable elements, rules of form, and content supporting both production and consumption - the basic determinants of a discourse genre. Reiffel (1999) describes how the highly stylized form of mathematics writing serves the community of scholars in this domain well. Conformance with genre conventions enhances memorability of discourse (van Dijk and Kintsch 1983) and leads to greater user satisfaction (Bazerman 1988). Researchers in the area of hypermedia and Web design have noted that user orientation and navigation is contingent on the user's perception of such rules in the information space and therefore the lack of genre conventions in the digital world is a potentially significant source of user difficulty. 8

The Web page as genre is still in formation. The guidelines and criteria that exist are minimal and often based on casual observations or library folklore. There are no laws and standards governing the information provided on Web pages and how it is organized. In addition, the genius of hypertext is that its texts are not necessarily linear nor is the text "fixed" or permanent in the way print technology fixed texts. In fact, it could be said that the nature of hypertext technology and the current organization of the Web are producing a standard of constant modification and reinvention for Web pages. The purpose of Web pages, the tasks for which they can be used, the ways users navigate the Web, the relationship between Web page design and Web page use or between page design and the information being represented are all current questions and problems for researchers and practitioners alike.

Nevertheless, there are developing standards that let us share the formats and codes necessary for access to material on the Web. 9 The Web page as genre is forming both within institutions and outside them. Within corporations, libraries, and universities, patterns of Web page design are developing, in part through the publication of how-to articles, workshops, and the growing professionalization of corporate Web page design. At the same time, the extraordinary and unprecedented ability for everyone to be authors, designers, and publishers means that a lot of Web page design is taking place outside these institutions. The shared uses to which these Web pages are put, such as the distribution of hobby or fan information, professional résumés, and family photos might help shape formats and expectations. Web page authoring software will probably also shape format, in part by providing templates. 10

University Home Page as Genre

Dillon and Gushrowski note that genres evolve from particular discourse communities. In this case, we are talking about colleges and universities and their participants. Given the nature of hypertext as a medium, university Web page designers find themselves working within a paradox: they manipulate a potentially fluid and nonlinear medium for complex, traditionally structured institutions. Unlike electronic library catalog records, which represent documents according to strict, shared, and pre-established formats and codes, there are no codes or generally accepted models for representing universities on their home pages.

Despite the absence of standards, an analysis of 111 university and college home pages reveals common design elements, rules of form, and content. Page layout is remarkably similar. All place the name of the institution as a banner across the top of the page. All but four (96.3 percent) of these home pages provide photographs of the most picturesque aspects of the campus. All provide keyword links (Admissions, Athletics, Student life, etc.) to university offices, activities, and services. The text, links, and photos below the banner are organized into columns. Another column on the left or right side of the page often displays another set of links. A look at the source coding for these pages indicates that the use of the tables function for layout in HTML is responsible for this common structure. There is frequently a navigation bar along the top or bottom of the banner with the university name providing an additional set of links, and almost always another with the links in text format rather than a graphics file at the bottom of the page. A growing number (currently eighty-three, or 74.7 percent) include news and headlines of activities, events, speeches, and research findings at the university. Sixty-eight percent (n=76) use overt symbols of the university, such as the university insignia, somewhere on the page. These patterns are so common those few pages that do not use this format are startling, as deviations from the rules of genre generally are.

Criteria for Analysis

This analysis resulted from the disruption of a user's expectations. The author of this article was studying examples of subject-related Web pages designed by university libraries. The search focused on the fellow institutions of the Committee on Institutional Cooperation, the CIC. On several occasions, the link from the university's home page to the library appeared to be missing. That is, frequent use of university home pages had led to an expectation of certain elements, and one of those elements was not there. Hypermedia researchers have noted that ease of navigation depends on the "the user's perception of such [genre] rules in the information space." 11 Without a direct link to the library, the user was jolted out of the task at hand and had to develop other strategies to search for the library before searching for library research guides.

The criterion used to categorize these home pages was the presence of a visible active direct link to the library Web page. Of the 111 home pages examined, sixty-eight (61.26 percent) of the ARL college and university home pages had visible active links that said library or libraries. One additional group provided a link that grouped the library with other resources or units (i.e., libraries and museums) and led to an intermediate page that had a direct link to the library (see for example, www.ua.edu). A second subgroup of these pages did not have a direct active link to the library but informed the user how to find the library: that is, by listing the library under the heading services (i.e., www.uga.edu). A total of thirteen home pages fell into either of these two groups, placing the library into a larger category with some other services or offices.

The university home pages that are categorized here as having no visible link to the library require the user to make some choice or perform some action before such a link becomes visible. So these pages either have no visible link to the library at all (i.e., www.brown.edu), reveal a link to the library if the user clicks on a drop-down menu box ( www.unl.edu or www.bu.edu), or indicate a pathway to a library link when the user runs the mouse over a broad heading ( www.utah.edu). The analysis of ARL-member home pages showed that twenty-seven of the 111 (24.3 percent) university and college home pages had no active visible link to the library.

Web Page Design and Analysis

Web page design is currently taught as a form of classification, in library terms, or from an audience perspective in media terms. Before designers sit down to code a page, they are supposed to sketch out an organization of the information they are going to provide. This classification starts with the definition of the purposes of the home page itself. Is a university home page an electronic promotional brochure? If it is, for what audience has it been designed? Possible audience groups include prospective students, alumni, donors, and prospective employees. Is the university home page the entry to online services for students, faculty, and staff? Is it meant to be the on ramp to online registration and advising, human resources, and library materials? What sorts of information do such uses and users require?

In addition to identifying common features of this developing genre, we can read the resulting home page, that is, examine the communicative strategies of the media text. How is the user addressed: as a tourist, student, or potential donor? What sorts of imagery are used? What demands are made on the user? What symbols are they expected to understand? What actions do they have to take to interact with the page?

The identification of the purpose of the Web page and of the audience helps decide what links and information should be displayed on the home page. It is important to note that designers find themselves balancing between conflicting demands. In particular, anyone creating a useful Web page has to find a balance between providing all the available information and creating a page so cluttered that it is difficult or unattractive to read and use.

Researchers in cognitive theory and library sciences argue that the processes of classification, and the classification systems that result, represent our theories or understandings of the nature of our universe, or even structure that world for us. 12 Since it is impossible to provide links to all university and college units and services on the home page, some classification of those units and services takes place when designers create pathways from the home page to those sites. That is, the Web page designers form broad categories and assign units and services to those categories. The categories on university and college Web pages analyzed here fall into three broad types: units of the university (colleges, libraries), function served by those units (admissions, registration), and the role of the user (prospective students, current students, faculty, alumni).

Thus it can be argued that the way a university classifies or locates its libraries within the array of university units indicates how the university administration thinks of its libraries. That is, the classification of the library by the university represents the perceived role and status of the library. Thus we are examining how a selection of U.S. university home pages represents the place, ontologically speaking, of their libraries. The argument here is that the designers and those who approved those designs created certain definitions of the library with their design of the home page: as either a student service along with parking and computing; as a principal unit of the university's academic purpose; or as part of the university's information infrastructure.

No claim is being made here that this ontological understanding of the library is systemic within each university. The representation of the library on the university home page may or may not reproduce the university's way of seeing the library when it comes to funding, faculty status for librarians, etc. Nor is it necessarily the case that the library has one and only one definition within the university or only one place on the university Web site. Nevertheless, we can read the Web page as one officially approved representation of the university's structure.

Page Design and the Placement of the Link to the Library

Organizing a representation of the university by its offices, schools, and departments is perhaps the most traditional choice. The university itself as an organization is structured into administrative units, colleges and schools (and those into departments and programs), and services. A home page organized that same way would appear to mirror the working organization of the university. So the home page of Penn State University ( www.psu.edu), for example, presents links that are the names of such units (academic programs, libraries, athletics). Each of these names functions as a heading, with the names of more specific units following. Under academic programs on the Penn State University page, for example, are live links labeled undergraduate programs, graduate programs, international programs, and more.

A variation of this organization scheme lists university units by their function. The Indiana University-Bloomington, page ( www.iub.edu) uses academics and research, admission and advising, student services, and libraries and computing, among others. The headings are followed by listings of specific units. The libraries and computing heading is followed by the terms "Catalog, databases, Telnet . . ." with the ellipsis indicating additional choice. These listings are not live links and the user must click on the heading, which leads to an intermediate page of links. In this case, the first link on the intermediate page is to the IUB Libraries. The other links on that intermediate page are to computing services and resources. This linking of the library and computing services represents a change in the schema of categories on this page. An earlier Indiana University page listed the library under the heading of student services, after parking and health services. Of the thirteen home pages that group the library with other units or services, seven now link the library to information technologies, six to museums and the arts, and three to research. 13

This design attempts to organize the links and control clutter while still providing information to the user. The words in the lists below the headings could of course be active links, saving the user another step in the path to the library. But the use of the ellipsis on the University of Illinois page allows the designer to use interior pages for long lists of links. The Penn State University page uses live links beneath the headings that, while creating a wordy and long page, also pinpoint the specific target: the user has the choice of going directly to the library home page, the catalog, the database search page, and others.

Another organizational schema used on university home pages is based on the functions performed by these units. Common links on these pages are admissions (the process rather than the office), research, learning, health care resources, computing, and giving. An organization purely by function might find it more difficult to conceptualize a place for a link to the library. Those pages often have links to the library on more than one intermediate page (i.e., to be found by clicking on research or learning).

The third schema, which has become more common over the duration of this project, is to organize information by the role the user plays in the university community. Common headings here are prospective students; current students; faculty and staff; and alumni, visitors, and friends. This can create a very clean design but one that provides very little information and no place for a visible library link. A number of these ( www.brown.edu, www.yale.edu, www.ucsb.edu) provide additional information through mouseovers, attempting to remove the clutter from the home page by having it pop up and then disappear as the user moves the mouse.

Most university home pages use at least two of these categories of links. Sometimes they are mixed together, while others clearly mark the different categories, i.e., information for (roles) and information about (units and functions) (i.e., www.ua.edu, www.umd.edu). Others put one set on the navigation bar and another in the middle of the page or in a column (i.e., www.psu.edu, www.ucsd.edu). A link placed on the bottom of a Web page, using whichever classification scheme, runs the risk of disappearing from the view of the user. Since browsers are configured differently on individual machines, any designer who places a link at the bottom of a Web page runs the risk of hiding that link if the user does not scroll. Years of research into print design principles argue that the location of material on a page should be related to importance, and that important material should lead.

Web Design as the Classification of Information

When forced into searching for the library on some of these sites, users have to play a sort of classification game: for those universities that did not have an active link marked library on their home page, which link on that page hid - or led to - the library link? Was it classified under academic units? Student services? Information resources? The answer varies widely.

The links on the Brown University home page ( www.brown.edu), for example, are arranged around the concept of activity, rather than university unit or the user's identity. Eight links are aligned to the right of a photograph: learning at Brown, research at Brown, working at Brown, before Brown, about Brown, after Brown, and Brown in the community. There is no other text below the banner. 14 The page has a very clear, simple design with a lot of white space. The banner itself contains text links on a navigation bar below the university's name: people and places, news and calendars, A to Z, search, and about this site. No single university unit (i.e., a college or office) has a link on the home page. The most consistent element on Brown's home page is in fact the repeated use of the word Brown, used ten times in that spare design. The banner graphic includes the word university, but the links do not necessarily define a university. Brown could be a museum, a nature conservancy, or an institute of any kind.

The Web visitor sees no indication of where to find the link to the library. As the user's mouse moves over each link, JavaScript-enabled text appears with further information about that heading or category. Thus moving the mouse over the learning at Brown link produces the caption academic departments, services, and rules. But none of those rollovers mentions the library. Using the click-on-everything method of searching, the user finds a link called libraries and museums on the page reached by clicking learning at Brown ( www.brown.edu/Webmaster/learning.html), information not provided by the rollover script. The rollovers solve the clutter problem, but they do not direct the Web visitor to the library.

Just in the time it took to write this article, Yale University has presented three different home pages. When first examined, Yale's Web home page was designed so that the links were placed on the images of old books, but ironically there was no visible library link. The click-on-all-links method found it under Academics ( www.yale.edu/acad.html). Only printing the page clued this user to the fact that there is a list of headings and captions lower down on the page that includes a library link. The user had to scroll to find them but the design of the page did not indicate more information follows. However, a note to visitors on the visible top half of this version of Yale's home page suggested that "frequent visitors might prefer to open to YaleInfo" ( www.yale.edu/yaleinfo). YaleInfo was an alternative home page, structured by such traditional headings as academic departments and schools, courses, library, computing, and student life. The link to the library was at the center of the YaleInfo page.

The current Yale page ( www.yale.edu) resembles Brown's page in design to the extent that it uses text sparingly and includes living at Yale and working at Yale as well as links with such titles as academics and alumni. Moving the cursor over these headings makes JavaScript-enabled subheadings appear. But the current Yale page communicates the link to the library more explicitly than the Brown page, since moving the mouse over the academics heading makes an active link to the library visible on the home page.

Using the Search Engine

Random clicking away at links is hardly a satisfying search strategy for a librarian. But while search engines are present on 96 percent of the university home pages examined (n=107), the results of searching for "library" are imprecise. An analysis of the home page of the University of Michigan at Ann Arbor ( www.umich.edu) can be taken as an illustration. U-M has a superb library, a top-ranked school of information sciences (which includes library science), and is at the forefront of university information technology and development. Until 2001, however, the university home page had no direct link to the library.

A look at the five most prominent links, centered just under the banner, indicates they are defined by user role, not by university units: prospective students;alumni and friends; visitors, scholars and researchers; and students, faculty and staff. Below that and to the right, we find five more links. Only one, corporate partners, fits into the user category. The others (academic units, athletics, U-M health resources) refer more traditionally to university units. The link community assistance fits into neither category and takes the user to a description of a special university program of community service. When a link to the library was added in 2001, it was added to this group.

Before the link was added, an attempt was made to find the library's home page by using the search engine. Entering the word "library" in U-M's search engine from its home page resulted in seventy-five hits (three printed pages of results). Scattered among the hits on the first and second Web pages are links to various specialized libraries in the U-M system and to MIRLYN, their online catalog. The link to the library's home page is found at the bottom of the second page of results. Going back to the search page to try to design a more precise query revealed drop-down menus lower down on the page that narrowed the search by subject or field. "Library" is listed on that menu as a subject, but it does not function as a link. The user must enter a keyword into the search and select the library as subject. Fifty-two hits result, and the library's home page is somewhere on the second page of the alphabetically arranged results.

Redundancy

Reverting to the click-on-all-links method, though, revealed that clicking on any of the links defined by user role leads to pages composed of long lists of links. Among those links is one to the library. The library link is actually most accessible from the alumni and friends page, where it appears near the bottom of the screen as soon as the page opens. The user has to scroll down to find the library link on the students, faculty, and staff page; the visitors page; and even the scholars and researchers page, where the library is at the end of a very long page of links, since the links are arranged alphabetically after a general information link. 15

The library was not listed on any of pages reached by the lower links: academic units, athletics, community assistance, corporate partners, or U-M health resources. The Web page for corporate partners ( www.corporaterelations.umich.edu) notes that its goal is to "outline the variety of ways in which corporations, foundations, and other organizations can enter into partnership with the University. These include the following: Colleges, Schools and Other UM Organizations, Recruiting, Research, Continuing Education, Academic Outreach, Diversity Programs [and] Technology Transfer." It seems a useful place to have a link to the library, which could (and does) play a significant role in most of the areas listed.

The important point here is that the designers of the University of Michigan gateway or home page originally chose not to put the library link on that page. But they did respond to the problem faced by the user ("Where is the link to the library?") through redundancy: every user, clicking on the link displaying his or her identity as visitor or student or faculty, will find a link to the library if they are willing to scroll down the page looking for it. Since the addition of the visible link to the library in the second column of links, this sort of cognitive identification of role is no longer a precondition to finding the library: the user, whoever he or she may be, can directly access the library Web pages.

Conclusion

Almost a quarter of the colleges and universities that belong to ARL do not provide an immediately visible active link to the library on the university home page. The new genre of university home page provides a number of common locations for that link: in the main columns of links, in navigation bars, in lists of Quick Links. And while the University of Michigan, Cornell University, and the University of Minnesota all added previously missing links to the library to their home pages this past year, the new page for Texas A&M University, which was mounted in February 2001, removed the library link. There is, however, a link for parking.

Another study could analyze the ways in which the library is categorized on those pages where it does not have its own link: whether it is considered part of academics, resources, services, museums, or learning. An unambiguous conclusion of this study is that the Web page designers of 24 percent of ARL member institutions do not consider the library a primary destination for their Web visitors.

It seems obvious - now that university and college library Web sites provide remote access to indexes and catalogs, full-text databases, electronic journals, and e-books - that such a service should be highlighted on the university's primary Web access page. A university interested in its presence in cyberspace should be celebrating its cyber libraries.

This study also documented the constant mutability of Web pages. During this project, the University of Minnesota, Arizona State University, Texas A&M University, the University of Chicago, Indiana University-Bloomington, MIT, and probably others redesigned their pages. Yale mounted three successive versions of its home page. Smaller changes abound, such as the addition of library links to the University of Michigan, Cornell University, and the University of Minnesota pages in the past year, and Brown University and Yale University's addition of announcements. These pages are continually tweaked, headlines are updated, links regrouped. Adding links to the library would be easy, useful, and logical.

That logic, though, might be challenged by the development of what might be a new genre of university home page design. As noted earlier, almost 75 percent of the home pages studied included headlines or news stories about their universities. A few have gone a step further, and have created Web pages that look like university magazines. Kent State University has placed all the role, units, and function links (including one to the library) on navigation bars that border the top and bottom of the page ( www.kent.edu). The rest of the page is designed as a magazine cover, offering tours, news, a welcome from the president, and a link to another magazine titled "Success @ Kent State." Southern Illinois University's page is even more like a magazine cover. A large photograph, a portrait, comprises most of the page. Superimposed on the photo are the headings news, cover story, and from the chancellor. A button on the lower navigation bar links to past stories. A kiosk to the right of the photo provides university links, though not one to the library.

The most overt example of university home page as public relations magazine is the home page of the University of Nebraska ( www.unl.edu/unlpub/index.shtml). The page provides the full text of several articles about the university and a column on the left side that offers links to videos of campus, a welcome from the chancellor, and a history of the university. The URL makes it clear that this is a publication of the university's public relations office. A navigation bar across the top of the page, designed to look like the tabs on file folders, leads to more traditional links (academics, students, admissions, etc.). There is no visible link to the library, though one can be found in a drop-down menu box. The title of the page that appears in the browser history is "There is No Place Like Nebraska" rather than the name of the university, as usual.

The interior pages of these sites include links to the libraries along the same patterns identified above. But if university home pages begin to cover only what look like press releases, users will find themselves yet another step away from online resources. Many of the home pages examined gave the weather report. They would better serve their users if they gave direct access to the library.



References and Notes

   1. For a complete list of ARL members, see www.arl.org. This study excluded specialized research institutes, and analyzed only university and college home pages. Two universities, the University of Arizona and Texas A&M University, were both previewing new home pages. Their existing and their new pages were analyzed, bringing the total number of home pages analyzed to 111.

   2. Andrew Dillon and Barbara Gushrowski, "Genres and the Web: Is the Personal Home Page the First Uniquely Digital Genre?" Accessed July 13, 1999, www-slis.lib.indiana.edu/adillon/genre.html; citing Andrew Dillon and Misha Vaughan, "'It's the Journey and the Destination': Shape and the Emergent Property of Genre in Evaluating Digital Documents," New Review of Multimedia and Hypermedia 3 (1997): 91-106, www.slis.lib. indiana.edu/adillon/NewReviewPaper.html; Tom Erikson, "Why the Future of the Internet Has More to Do with Genre Blending than Gender Bending," Accessed Dec. 15, 1997, www.pliant.org/personal/Tom_Erickson/genreBlending.html.

   3. For a useful summary of the history and variations of genre theory, see "Genre," in Key Concepts in Communication and Cultural Studies, 2d ed., ed. Tim O'Sullivan et al. (London: Routledge, 1994) and Daniel Chandler, "An Introduction to Genre Theory," Media and Communication Studies Site, Nov. 8, 1997. Accessed July 13, 1999, www.aber.ac.uk/media/index.html.

   4. Denis McQuail, Mass Communication Theory: An Introduction, 2d ed. (London: Sage, 1989), quoted in Chandler.

   5. Todd Gitlin, Inside Prime Time (New York: Pantheon, 1985).

   6. The term "recombinant genres" is Gitlin's.

   7. Elizabeth Eisenstein, The Printing Press as an Agent of Change, Communications, and Cultural Transformations in Early Modern Europe (Cambridge Univ. Pr., 1979.)

   8. Dillon and Gushrowski. Their citations are to E. Reiffel, "The Genre of Mathematics Writing and It's [Sic] Implications for Digital Documents," in Proceedings of the 32d Annual Hawaii International Conference on System Sciences (Los Alamitos, Calif.: IEEE Computer Society, published on CD-ROM); Teun van Dijk and W. Kintsch, Strategies of Discourse Comprehension (London: Academic Press, 1983); Charles Bazerman, "Shaping Written Knowledge," in The Genre and Activity of the Experimental Article in Science (Madison, Wisc.: Univ. of Wisconsin Pr., 1988); and Dillon and Vaughan.

   9. John D'Angelo and Sherry K. Little, "Successful Web Pages: What Are They and Do They Exist?" Information Technology and Libraries 17 (June 1998): 71-81.

   10. Henry Jenkins, Textual Poachers: Television Fans and Participatory Culture (New York: Routledge, 1992).

   11. Dillon and Gushrowski (see note 2), citing Dillon and Vaughan.

   12. George Lakoff , Women, Fire, and Dangerous Things: What Categories Reveal about the Mind (Chicago: Univ. of Chicago Pr., 1987); Roberto Poli, "Ontology for Knowledge Organization" in Advances in Knowledge Organization 5 (1996): 313-19; Michel Foucault , The Order of Things: An Archaeology of the Human Sciences (New York: Pantheon, 1970); Hope Alene Olson, "Dewey Thinks Therefore He Is: The Epistemic Stance of Dewey and DDC," in Knowledge Organization and Change: Proceedings of the Fourth International ISKO Conference 15-18 July 1996, Rebecca Green, ed. (Frankfort/Main: Indeks Verlag, 1996), 300-12.

   13. The number adds up to more than thirteen because the groupings are sometimes larger than two. The University of Rochester (www.rochester.edu) has a link on its navigation bar to "Libraries, Info Technology, Art Museum" and the University of Southern California has a link to "Research, Libraries, Computing."

   14. This is usually the case. When the search for a new university president started, a headline about the search was added below the photo. Now that a new president has been chosen, the date of her inauguration was announced under the photo.

   15. As a reviewer pointed out, whether a user has to scroll to see an element on the page depends on the browser and the monitor settings. In this case however, the "print preview" function of the browser indicated four printed pages of links, with the link to the library at the bottom.


   Roberta Astroff is Humanities Librarian and Coordinator, Digital Resources Center, Arts and Humanities Library, Penn State University.