In Elastic Windows, hierarchical information is mapped onto nested
rectangular windows. In this mapping, higher-level information
encapsulate lower-levels within its window borders. For example,
assume you, as a professor at the University of Maryland (UMD), are
organizing your e-mail messages according to the academic institution
hierarchy at the university (Figure ). In this,
example, your e-mail messages are classified into two, College of
Computing Mathematical and Physical Sciences (CMPS) and School of
Engineering (ENG) below the top-level (UMD). Messages are further
classified according to departments, and then by the occupation of
each sender. In this example, non-leaf elements of the hierarchy
solely function as grouping elements. They do not contain information
by themselves.
Consider another example, where you are browsing the WWW page
documents at the University of Maryland site
(Figure ). In this example, the UMD main web page
document is visited initially, and then the CMPS and ENG web pages are
visited subsequently. Next, departmental web pages are followed for
each school. Finally, web pages for students, faculty and staff are
browsed for the department of Computer Science (CS). In this example,
non-leaf elements of the hierarchy also contain information to be
displayed. According to the mapping in this example, the window of the
linked child document is nested within its parent window next to the
parent document. Alternatively, the windows of each child document can
be allocated space below the parent window side by side.
Representing hierarchical information by mapping onto nested windows has many benefits. Firstly, nested windows achieve space scalability. As more and more information is included in the information space, the screen space requirements does not increase linearly since screen space for lower-level windows is bounded by the space allocated for the higher-level window. However, screen space allocated for each new window gets smaller as new windows are inserted.
Secondly, the hierarchical nesting clearly depicts the semantic relationships that exist among pieces of information in the hierarchy. Nesting is considered to be a very good visual presentation of the inclusion relationship. Besides, nesting also facilitates a visual abstraction mechanism, where lower-level information is visually included in the higher-level windows.
Finally, nested windows provide a visualization of detail within an overview as opposed to detail and overview separately. This eliminates problems due to divided attention of users and also allows users to build up spatial memory more easily, since the overview is constructed directly by the user in the same screen space.
In terms of space utilization, depending on the size of information at each level of the hierarchy, there may be empty regions at higher-level windows when the information in these windows falls short of the window height. Typically, information at each level is long enough to fill the whole contents of the window.
Another disadvantage of representing hierarchies by nested windows is that less space is allocated to lower-level information in the hierarchy. This could be a problem when information in lower-level windows does not fit into small windows.