Go to previous article
Go to next article
Return to 2005 Table of Contents
Presenter(s)
Jesse Walker
Sarah Swierenga Email: accessibility@udri.udayton.edu
Laura Militello
Megan Gorman
University of Dayton Research Institute
Michigan State University Email: uac@msu.edu
Sightless users, as well as those with low vision, continue to struggle to interact effectively and efficiently with a range of websites. The lack of cues available to sightless users coupled with the serial nature in which they are presented using screen reader technology, e.g., JAWS for Windows, or text linearization software applications, e.g., Delorie's Lynx Viewer, often make it difficult to find and follow a path that leads to task completion (Swierenga, Militello, Gorman, & Walker, 2004; Gorman, Militello, Swierenga, & Walker, 2004). Further, after a sightless user has begun traversing an incorrect path it is often difficult to determine how and when the error was made and what should be done to rectify the error in order to continue along the correct path. Clearly, mechanisms to aid the sightless user in finding and following a path that will lead to task completion are needed. Websites tailored specifically for sightless users would clearly identify core functionality first, so there would be less ambiguity about where on the page the starting point could be found. Rather than relying on cues such as font color and size or visual spacing to guide the users, a more hierarchical presentation of information based on conceptual models that sightless users form of website structure would allow them to "scan," or listen to, major functionality or information content areas. Users could then deepen the one that seems most promising, rather than listening through a sea of undifferentiated links, graphics, text, and tables, with no cues as to what the designers thought worthy of their attention.
Enterprising individuals and groups have begun this process of making various aspects of the web more usable to more users. Locke, et al. (2004) have addressed automating internet tasks where there are clearly defined steps and goals that a user may utilize and adopt when interacting with a website. Their proposal assists users where a clearly defined goal exists (e.g., install software), and where the primary task is comprised of a limited number of subtasks. The user may choose to engage in any of these subtasks based upon his or her expertise, but the end goal remains consistent. In fact, their proposal allows automation or semi-automation based on the presence of this well-defined goal. From another standpoint, Mayer & Bederson (2001) have worked with methods to allow web users to track paths they have previously taken. When an individual completes a complex search and finds the intended destination, BrowsingIcons and BrowsingGraphs help the user map that destination for future use. These visualizations serve as an atlas, allowing the user to jump in at any point of the mapped browsing session.
Against this back drop, usability and accessibility specialists from two Midwestern universities are exploring the development of an XHML applet tool, called the Task-Based User Interface Filter Method and Apparatus (2004). An XML-based language, called TaskML, is also being developed to program the user interface. The task-based interface tool will allow users to interact with data by navigating websites in a hierarchical, context-driven, task-based mode. It will provide a framework for guiding users through uncharted information or web pages where the goal is ambiguous, changing, or very loosely defined. Unlike the Locke proposal (2004), which addresses specific, linear task processes (e.g., install software), and thus is highly linear, or Mayer & Bederson's Browsing Icons/BrowsingGraphs concept (2001), which has a heavy visual-spatial focus, the current task-based interface tool will support users, who have changing or ambiguous goals when they visit web pages or interact with data. Users may initially arrive at a page seeking, simply, "information." The page may provide many different types of information, and many different means to arrive at this information.
The tool will support users attempting to sift through the varied and frequently confusing interface, even while they shift from seeking information, to providing information, to completing an activity. To use this tool a sightless user might visit weather.com to find the current local temperature. A pseudo dialog box would list the key tasks available on the home page in the website, as well as the core global tasks on the website. Task descriptions would also be included to enable the user to be fully informed before selecting a task. The task list would be regenerated each time a new page came up. Each new page would surface the next step in the process, highlighting what the user needs to do to continue down that path. A small set of core functions would be read (perhaps as links). The user would select the find temperature functionality and immediately be presented with the edit box asking him or her to enter a zip code or city name. Having entered this, the user would be immediately directed to the results of the search task (instead of being introduced to the top of the results page, which is an ambiguous cue due to the repetitive nature of the navigation and advertising links). As the sightless user became more advanced, he or she would be able to discover new features and tasks that would complement his or her level of expertise. In summary, developing research-based accessibility tools would enhance website efficiency, usability, and accessibility for both sightless and sighted users.
The task-based interface tool will provide a more meaningful, effective and efficient experience when interacting with complex sets of data, large web sites, web applications, and databases by integrating a task-based view of the data set into existing and new user interfaces for the web. The tool offers a way to provide users with the user interface architecture, surfacing the design of the site. Alternatively, it separates the content of the data from the presentation and/or information architecture of that data. This separation is similar to the effect that is currently found by using cascading style sheets in conjunction with HTML-based web pages.
This task-based interface tool provides advantages for content developers as well as end users. The task-based interface filter provides a framework that allows them to focus on the data sets and the tasks associated with the data set without having to design an interaction scheme that may or may not allow meaningful navigation and manipulation of said data set. This allows the developers to separate content from presentation making the data sets more flexible. This could also potentially result in faster development and deployment of content since the developers are working within a consistent framework.
The advantage to the end user is also significant, especially users who rely on screen reader assistive technology. An end user would be able to choose a particular data set and be provided with a simple hierarchical method by which to determine how the data set can be navigated and manipulated. Since the tool provides a consistent interface in which to interact with data, the user does not have to decipher a new interface scheme each time he or she is presented with a new data set. Instead, the user can interact with this new information in a single window. While providing a consistent interface regardless of the data set being accessed, the task-based interaction can more quickly allow an end user to discern what actions can be taken with the chosen data set. Providing a predictable, context-driven, task-based interface will also allow the end user to develop robust mental models of the data sets more quickly, making data more accessible to computer system users. Placing the responsibility of the management of the tasks involved with a particular data set on the computer system allows the user to focus the goal of the interaction rather than spending time interpreting the possible tasks based on an inconsistent interface. In summary, the tool will increase accessibility for all user groups, including sightless users, low vision users, inexperienced computer users, and users dealing with complex web applications and data sets.
References
Gorman, M., Militello, L. G., Swierenga, S. J., & Walker, J. L. (2004). Internet searching by ear: Decision flow diagrams for sightless internet users. In Proceedings of the Human Factors and Ergonomics Society 48th Annual Meeting. Santa Monica, CA: Human Factors & Ergonomics Society, 243-247.
Locke, B. S., Breck, G. M., Brousseau, D. A., Fitzpatrick Jr., R. S., Playdon, P., Tran, K. H., & White, K. A. (June 3, 2004). Method and apparatus for navigating through a task on a computer. U.S. Patent Application Publication, No. 2004014939.
Mayer, M. & Bederson, B. B. (2001). Browsing Icons: A TaskBased Approach for a Visual Web History, Maryland: University of Maryland, HCI Lab. HCIL-200119, CS-TR-4308, UMIACS-TR-2001-85. http://citeseer.ist.psu.edu/mayer01browsing.html
Swierenga, S. J., Militello, L. G. & Gorman, M. (2004). "Going in blind doesn't help": Cues for navigation. In Usability Professionals' Association 2005 Conference Proceedings. Bloomingdale, IL: Usability Professionals' Association.
Walker, J. L., Swierenga, S. J., Militello, L. G., Gorman, M. E. (filed October 1, 2004). Task-based user interface filter method and apparatus. U.S. Patent Application Publication Pending.
Go to previous article
Go to next article
Return to 2005 Table of Contents