The Temporal Organisation of Documents and Versions: A User-Centred Investigation
Thesis DisciplineComputer Science
Degree GrantorUniversity of Canterbury
Degree NameDoctor of Philosophy
In this thesis a study of computing systems that use time as the primary method of organising electronic documents, and versions of electronic documents, is presented. Such systems should be useful and usable because they exploit people's intuitive understanding of temporal order. In addition, the systems are worthy of investigation because they have received little attention, yet they may provide enormous benefits to users with little cost, as the temporal information is easy for systems to record. Temporal document organisation systems have the potential of alleviating many of the problems that traditional systems have had for decades. Throughout this thesis, user-interface guidelines for the implementation of temporal document-organisation systems are presented. The guidelines are based on empirical and theoretical evaluations that I have conducted, and studies of other's work. By using theses guidelines, designers should be able to create interfaces that are liked by users, and provide good support for the user's tasks. The first set of guidelines are based on an investigation of the human factors of temporal document-organisation, specifically looking at memory and temporal awareness. These human factors are related to the user's tasks with documents and document versions: finding, reminding, error-recovery and system exploration. Another set of guidelines, based on a study of how existing document-organisation systems support the user's tasks, are then presented. My first empirical evaluation looks at history lists, which are some of the most common temporal document-organisation interfaces that are found today. In the study it was found that participants are slower at retrieving Web pages when using an interface that broke the history into non-temporal categories than with the other three interfaces that were tested. In addition the participants preferred the interface that broke the history into 'temporal chunks'. Following on from the history-list evaluation, a theoretical and empirical evaluation of version retrieval systems, including undo, is presented. It was found that, in a text-editing environment, there is sufficient mechanical reason son for forward error-correction to be favoured over undo when correcting small and simple errors. For more complex errors, it was found that a visualisation of the prior document versions is better than forward error-correction and undo. In a similar evaluation of error recovery in a drawing editor, undo was found to be the quickest method of recovering from simple errors, while a visualisation of the prior document versions allowed for faster recovery from more complex errors. Having looked at the retrieval of documents and document-versions separately, my final study looks at a system that combines them both. The system that I developed organises documents without the need for file names and folders, which are used in most document organisation systems. In the formative study I found that the system that combines the retrieval of documents and document versions is useful and usable, and the organisation of the data did not confuse the participants. After each of the evaluations, I provide guidelines that should be applicable not only to the interfaces that were studied, but to temporal document organisation interfaces in general.