We choose to model our experimental task after one of the most frequent interaction tasks today's computer users do: Web browsing.
The task presents a total of 10 HTML pages to the subject, one page at a time. The document contains texts from an IBM computing terminology dictionary. A hyper link is imbedded at an unpredictable location in each page. The user's task was to scroll the document until he/she found the target hyperlink. Clicking on the target word "Next" would bring the subject to the beginning of the next web page.
Two random characters are presented together with the target hyperlink, in order to test mixed scroll/typing/pointing all at once in the future. Subjects can be required to type the two random letters into another window. This feature has NOT been used in our current study.
The size of the web pages was set as such that the scroll handle was 1.3 cm (so it was not too difficult to acquire for the standard mouse condition). The web browser viewing area was 24 cm wide and 15 cm long on a CRT display, which is the full size of a IBM ThinkPad display. Note that the size of the viewing window will affect the total completion time of the task. Make sure you set the window at 24cm by 15cm so your data can be compared against ours.
To avoid transmmission delay in the experiment, cache all 10 pages into memory first so the loading time of each page is minimal.
Performance timing is done by a Java appelet written by Barton Smith. To see (or copy) the timing results, activate "show Java console" from the option menu if you run a Netscape Navigator. If your browser does not have a Java Console (such as Microsoft Internet explorer), run a stopwatch program to do the timing.