WebTOC is one of the few web site visualization tools currently available (others are Mukherjea's NVB, and Keith Andrews' Hyper-G). WebTOC differs from these tools in that it has a table of contents format. The advantage of a table of content over other types of visualization is that users are familiar with its analogy to table of contents in books. Contrary to expectations, our results showed no significant difference in task performance times for all of our tasks. However, our results indicate that users prefered having a table of content for a web site regardless of performance. Having an overview can guide the user's decision making as to which paths they can follow to find their target page.
Since the purpose of our experiment was to compare WebTOC, Textual TOC, and Netscape, we had to confine our tasks to those which can be performed by users using all three interfaces reasonably. Consequently, we were not able to fully test the usefulness of the size and color coding capabilities in WebTOC. One of the capabilities we did not look at is a compact overview of the size and data format of the content of a web site. This compact overview is useful in identifying aggregate information for the site. Web users can make decisions on whether to browse a web site by looking at this overview. For example, if they were looking for a site with images, they could quickly eliminate a site without browsing through the pages of the site by simply looking at the overview. This feature is especially useful to webmasters. The size and format information can be used to guide them in organizing and managing their web sites. Highly unbalanced branches of the content hierarchy of web sites can be easily identified. This task would be difficult to perform by browsing links or following directory structures.
In addition to web site use, the application of WebTOC on cleaner hierarchies (e.g. UNIX directory structure) could be tested. In such a study, the effect of reading web pages, loss of context of link texts, highly unbalanced hierarchies and non-hierarchy architectures of web sites could be eliminated. The provision of size bars, flexible control of the hierarchy, and usage of the search option makes browsing hierarchies much easier, WebTOC could be tested against other hierarchical browsing techniques.
This is the first usability study conducted for WebTOC. The scope of the study was relatively small and more subjects should be tested in order to achieve more significant results. We recommend that this experiment be re-run with at least twenty subjects per treatment. We also recommend careful selection of the site to ensure that it has a richness of content and a suitable hierarchical structure. Finally, we recommend that the tasks be carefully examined to make sure that they will not lead to some of the same pitfalls experienced in this study.
This study is by no means a comprehensive test on all the capabilities of WebTOC. It primarilly investigated the usability of WebTOC as a navigational tool for users. Some other aspects of WebTOC that should be examined are:
User comments, our own experience using WebTOC and our observations of subjects have led to the following suggestions. Some of these recommendations were already integrated into WebTOC but were not activated in the version used in our experiment. One of the most common requests was for an option to search the hierarchy and present results in the context of the structure. This option is now available in WebTOC Other users requested more meaningful labels in the hierarchy. This could be achieved by either allowing the user to edit the entries or by using titles of pages instead of the text of the links. To increase the power of WebTOC as a browsing tool, history and user location tracking capabilities can be added.
Web sites are usually not organized in hierarchies, instead they often take the form of networks building on the capabilities of hypertext. Most websites are organized hierarchically on higher levels and becomes more complex when approaching the lower levels. WebTOC provides an overview that users would not get otherwise. The distribution and quantity of content is displayed with coding and grouping that can facilitate the user's decision making process. Although our results did not show statistically significant differences, our subjects showed a preference for WebTOC. Further study is needed before the strengths and weaknesses of WebTOC can be fully determined. We believe that with further usability studies and refinements, WebTOC will be recognized as a powerful browsing interface.