Personal information organization and re-access in computer folders: an empirical study of information workers

摘要

The current hierarchical folder system has long been found limited causing various difficulties in organizing and re-finding information on personal computers. Many alternative prototypes have been proposed to replace the current folder system. However, past empirical studies consistently observed that people prefer browsing folders in re-accessing information and only use searching as the last resort. Recognizing the complexity and our limited understanding of personal information organization and retrieval behavior in computer folders, my study was aimed to explore what people need from folders and the affordances and limitations of folders in the different stages of organization and retrieval, and furthermore provide implications for system design. Improved understanding on personal information organization and retrieval on computers is especially important today when personal information management (PIM) has entered public domain and the boundary between personal information management systems and general information systems becomes blurred. This multiple-case study investigated the participants??? information organization and retrieval behavior in their computer folders at four stages: keeping and discarding, organizing, re-organizing, and re-accessing. The difficulties they had at each stage were identified, and their computer folder structures and contents were analyzed. The participants include six PhD students and six administrative staff in an academic institution, with the former group representing the ???research??? end and the other one close to the ???administrative??? end in the activity spectrum proposed in (Bondarenko & Janssen, 2005). The data collection instruments include two rounds of in-depth semi-structured interviews, information re-access task observations, disk scans of several folders, and emails reporting re-access difficulties. The use of the two distinct groups of participants together with the multiple data sources and data collection methods provided rich and varied data for exploration and at the same time increased the opportunity to do triangulation in data analysis.The specific research questions include: 1) how do people keep and discard information items on computers, and what are on some people???s ???messy??? computer Desktops or in some ???messy??? folders? 2) From an integrative view, how do people organize information in computer folders, and what difficulties do they have in doing this? 3) What are the folder structures and contents like? 4) How do people re-organize folder structure and what difficulties do they have in doing this? 5) What are the tasks and strategies of re-accessing information on personal computers, and what difficulties do they have in re-accessing information? The result implies that: 1) people need an in-between mechanism for keeping or not keeping, as well as for discarding or not discarding a particular information item; 2) behind all the idiosyncratic folder creation behaviors, four elements are identified in how the participants organize information based on Hj??rland's typology of four views: rationalism, empiricism, pragmatism, and historicism; 3) the study identified two extreme types of folders in a spectrum ??? ???genre folders??? with no interfile relationships and ???project folders??? with complicated relationships between files, which can be partly attributed to the impact of using folders as workplaces. With all the three types of relationships identified in PREMIS observed in various ???project folders??? or folders close to this end, this study found that the various derivative relationships between files and/or groups of files led to the greatest difficulties for participants in finding and identifying files; 4) Behind the general browsing and search behaviors in folders, this study observed a re-access strategy similar to faceted navigation. The content analysis of the re-access difficulty examples reveals that the four FRBR tasks (finding, identifying, selecting and obtaining) in searching and making use of bibliographies and library catalogues also exist in information re-accessing on personal computers. This helps to define the ???re-access??? behavior on personal computers, and suggests the importance of the other tasks in addition to ???re-finding??? information.