Chris Dent
December 4th, 2001
The Indiana University School of Law has had a strong presence on the World Wide Web since very early in the history of the Web. The school developed their web site in 1994[1] and they appear to have been growing ever since. Despite the long history the site has several of the architectural problems that are endemic to the Web. I evaluated the site using Internet Explorer version 6.0. Casual investigation with other browsers indicates my experience may well have been different with a different browser.
The Law School web site is presented, for the most part, as two frames. The left frame includes:
· A navigational menu that hopes to indicate the structure of the site
· Links considered important
· A search interface
· A link to the parent organization, Indiana University
The right frame includes the current content under consideration. The home page includes:
· A feature story
· A picture of the law school
· A list of upcoming events
· Short notes of import
The two frames together set up the structural expectations of the user.
At the bottom of the home page there is a link, “Viewing this Site”. I hoped this would be instruction on how to get around the website. It is not; it is instead some information on what browsers work well with the site. Such is the state of the Web.
What follows is an evaluation of the web site based on its organizational structure, access points and navigational structure as recommended in class.
A website’s structure should be immediately apparent from the home page to set the stage for further navigation. The structure provides user pathways that, as Gullikson et al mention, “are expressed by categorical menu structures which contain four characteristics:…scheme…, categories…, labeling…[and] presentation.”
A scheme is the organizational approach used for creating the categories in the architecture. The Gullikson work lists seven common approaches to categorizing:
1. By semantic topic
2. By organizational structure
3. By spatial location
4. By chronology
5. By function
6. By user group
7. By frequency of use
Reiss focuses more directly on semantic and functional methods but agrees that category schema are the foundation of a good architecture.
The law school menu structure is hard to decode. It appears to be trying to satisfy too many users at once (suggesting a high level user group division may be in order). It is closest to a topical scheme, most notably because it lacks the verbs that Reiss believes go along with a functional scheme.
The menu labels do “avoid secret languages” (Reiss) and thus are probably meaningful to a wide audience. Following a menu item results in a reasonable page. Randomly selected search tasks such as finding a faculty or course description are fruitful.
This success, though, does not overcome a compelling sense of navigational confusion and ennui. No direction is provided by the topical scheme, again suggesting that a user group or functional scheme may be in order.
“Categories defined within the scheme must be distinct and mutually exclusive. This is particularly important as this determines the pathway that users take.” (Gullikson) This, perhaps, is the source of the navigational confusion. While I can find course descriptions successfully, I find them in at least two different sections of the web site. Once there, I don’t know if information nearby is related to courses or something else. There’s no APUPA.
As already stated, given the chosen scheme, the menu labels are fair. Their presentation, though, leaves something to be desired: Many are longer than one line. Is that two entries or one?
“The menu structure must be broad and shallow.” (Gullikson) This is somewhat true for this web site. It is broad: there are 20 links in the navigation frame. However, the structure completely falls apart when a user travels any depth. The hierarchy is clear to only one level. This makes navigation within the branches of the site confusing.
Access points are the tools users use to get into a site. They are the users point of entry in general and point of departure for specific information needs. Gullikson et al recommend several types of access points: multiple access points, a site map, an index, alternative menu structures and a search engine.
In addition to the primary access points of the left side frame and the search engine, the site does provide a sort of half-hearted attempt at an introductory access point in the “Welcome” document. There is a problem with this page though: choosing some links on it replaces the entire page with the requested page, destroying any navigational context the user might have had.
The site has neither map nor index that I could find, despite the value users find in such things. However, while searching for both the top most search result is documentation for web server administration.
I wish!
The web site does provide a quite complete search engine interface. The search dialog is available on all pages of the site that include the left hand side of the main frame. Exceptions include the Career Services sub-site and some of the sub-sites reachable through the Web Initiatives page.
There are, however problems with the search system:
· There is a huge volume content on the site thus result sets are very large
· When a link is followed the user is dropped into an uncharted sea
· Because frames are being used there is no visible URL for the produced page
· There are few navigational cues on the produced pages to let the user know where they are in the site structure and how to get anywhere else from where there.
· There are no indicators to suggest how a produced document is categorized in the architecture
Gullikson and crew encourage sites to have FAQs that users can use as an access point: “Users arrive at a web site with an expectation of finding one.” This site does not have a FAQ. It needs one.
Gullikson et al recommend three functions that should be on each page to enhance the navigability of the site:
· The ability to get to the top-level menu from anywhere on the site (including from sub-sites)
· The ability to use/access any of the access tools from anywhere on the web site
· The ability to determine easily one’s location within the hierarchy
Much of the time a user can reach the top level of the site if the left side frame is available. However as mentioned above the label is not very helpful. Clicking on “Law” does not mean “Home”, “Top” or even “Law School” if one wanted an explicit label. The same is true for the one access tool, the search engine interface. Both of these abilities go away, though, when the frame based navigation breaks.
Determining ones location within the hierarchy is difficult. Second level pages identify themselves with headers but third level pages are lost in the clutter. Follow the “Student Services” menu item for an example.
There is one primary recommendation for the Law School website: restructure the site to use a top-level user group oriented scheme with an underlying functional scheme to create user group oriented sub-sites. Such a restructuring will require several other adjustments:
· A site map and index should be created to augment the search facilities to allow users to browse outside the constraints of their user group.
· Trash the frame based navigation in favor of a template system that indicates more accurately where in the structure a user is located. Top-level structural features should be present on the pages at all time, but secondary structure should change depending where the user is in the site. This will allow the site to have more navigational depth while still maintaining a sense of space.
· At least two FAQs should be provided in FAQ repository that is reachable from all pages:
o A “How do I use this web site” FAQ
o An “Indiana University School of Law” FAQ. This could be potentially further divided into user group based FAQs
· The “Home” method should be more clearly distinguished from other menu selections as well as separated from the “Indiana University” link.
· If a user group oriented scheme is chosen, time spent identifying the various user groups and determining their needs is time well spent. The groups and needs will define the hybrid user group and topical scheme.
· Any architecture chosen must be maintained. A website cannot be left to sit. User analysis is an ongoing task.
Creating and maintaining an academic web site is a difficult task: the site has many different needs to satisfy. The different stakeholders of the organization frequently prioritize those needs in different ways. Is this site a marketing tool? Is it a student resource? Is it a faculty resource? Being willing to ask and answer these questions is the key to success.[2]
Gullikson, S., Blades, R., Bragdon, M. McKibbon, S., Sparling, M., and Toms, E. (1998). The impact of information architecture on academic web site usability. The Electronic Library 17 (5), 293-304.
Reiss, E. L. (2000). Practical information architecture: a hands on approach to structuring successful web sites. Harlow, UK: Addison-Wesley.
This project was assisted by a Java based web-mapping program of my own design, Webmap created as substitute homework for my CSCI A592 class. It can be used to create a hierarchal diagram of a website. Using it with the Law school’s web site was somewhat hindered by the fact that their web server is apparently doing something weird as a connection is made. This prevents a direct mapping of the server. Instead the two frame sources for the top of the site must be used to create two separate maps. This is not happening because of the use of frames. It is very odd.
The web-mapping tool was of little help in decoding the structure of the site because many of the pages have such an enormous number of links, even second level pages. The two maps may be viewed at http://www.burningchrome.com/~cdent/597/. These maps are five levels deep. Other depths of maps were prepared as well.
A graphical and interactive visualization version of Webmap is in the works.
[1] I rejected a job to be the Law School’s webmaster in early ’95. At that time their web site was already quite well developed.
[2] My redesign of the Honors College web site was hampered to the point of failure by a complete lack of willingness on the part of the administration to define the goals and users of the site.