Commercial Web browsers use a stack-based model for navigation. An alternative solution to the navigation problem is the temporal model [1], which maintains a complete list of previously visited pages. This paper describes an evaluation of these two schemes.
Browsing tasks that were evaluated include: hub and spoke navigation; spoke revisitation; parent revisitation; depth-first search and back; depth-first forward revisitation; within-site distant revisitation; and cross-site distant revisitation. The results showed that the relative efficiency of each scheme differed across different evaluation tasks. The temporal scheme supported backtracking to parent pages poorly, but performed better than the stack-based scheme for more distant navigation tasks. The temporal scheme also caused extreme usage patterns, with subjects solving tasks very efficiently or very inefficiently, depending on whether they used the back button.
The observation that subjects using the back button with the temporal scheme were particularly efficient led the authors to conclude that Web navigation could be improved by promoting this model. Currently, the authors are developing prototype systems that provide this functionality.