Towards Minimising Timestamp Usage In Application Software

With digitisation, work environments are becoming more digitally integrated. As a result, work steps are digitally recorded and therefore can be analysed more easily. This is especially true for office workers that use centralised collaboration and communication software, such as cloud-based office suites and groupware. To protect employees against curious employers that mine their personal data for potentially discriminating business metrics, software designers should reduce the amount of gathered data to a necessary minimum. Finding more data-minimal designs for software is highly application-specific and requires a detailed understanding of the purposes for which a category of data is used. To the best of our knowledge, we are the first to investigate the usage of timestamps in application software regarding their potential for data minimisation. We conducted a source code analysis of Mattermost, a popular communication software for teams. We identified 47 user-related timestamps. About half of those are collected but never used and only 5 are visible to the user. For those timestamps that are used, we propose alternative design patterns that require significantly reduced timestamp resolutions or operate on simple enumerations. We found that more than half of the usage instances can be realised without any timestamps. Our analysis suggests that developers routinely integrate timestamps into data models without prior critical evaluation of their necessity, thereby negatively impacting user privacy. Therefore, we see the need to raise awareness and to promote more privacy-preserving design alternatives such as those presented in this paper.