There are several impacts of our findings in this study. For designers of web sites, it might be best to make "index" and "table of contents" pages as dense as possible, the best case being for these pages to fit into one screen. These are the pages that need to be searched for the links of interest. Pages that are to be read more thoroughly can be made much longer and less dense because scrolling will not detract too much from the user's experience. We believe that the task of locating a link on a web page can be generalized to locating any item(something that can be selected) in a window of any application. This means that for multiple window applications, windows with lots of items should be made larger and their density increased so that more information can be presented at once. There is a balance that needs to be found in every application between the size of these more functional windows and other usable screen area for menus or other windows with textual and image information. The fact that a few people preferred smaller windows suggests that the windows in applications should be customizable. There are also implications for modern operating systems. Practitioners should look for easier ways to view file and application windows in graphical user interfaces. A user should be able to view several windows at the same time without having to manually adjust their size and make them harder to deal with due to increased scrolling. Perhaps some sort of dynamic zooming of the windows would be a desirable feature.
Future research should validate our findings in applications other than web browsers. Because of the trend towards multiple window applications, a goal should be to determine a good guideline for balance between the sizes of various types of windows. For example, a word processing program might have a main window for the document being currently worked on and several "toolbars" with icons to invoke various functions. Popular word processors like Microsoft Word and WordPerfect have implemented such functionality. The more toolbars that are place on the screen the less viewing area there is for the text. Research should determine how much space the menu bar and various toolbars can take up before productivity is limited due to too much scrolling. Another issue in this application area is the "floating toolbar" or toolbox which is not docked under the menu bar, but enclosed in a window. These windows often have to be moved around to uncover the working area underneath them or the working area must be scrolled past them to view the information that was hidden. Studies to determine how different configurations of the size of the main work area, toolboxes, and toolbars affect performance and subjective satisfaction would be useful.
A hypothesis is that each person has their own fairly constant variability between their performance times on the differently sized windows. This variability is perhaps caused by their method of scanning the screen for the target. We also believe that the time to move the pointer to the scroll bar and to scroll the contents of the window up or down is fairly constant for each person. Therefore, the main difference in performance times for each person in differently sized windows should be closely related to the number of times they had to scroll. Further research is needed to support or refute this hypothesis and perhaps discover mathematical methods for determining the relative constants for each person and predicting their performance based on the window size.
Continue
Return to the Title Page for Screen Size and Web Browsing