MultiLog : a tool for the control and output merging of multiple logging applications

Woodruff, Jonathan and Alexander, Jason (2016) MultiLog : a tool for the control and output merging of multiple logging applications. Behavior Research Methods, 48 (4). pp. 1296-1307. ISSN 1554-351X

PDF (MultiLog - for Richard Ingham)
MultiLog_for_Richard_Ingham.pdf - Accepted Version
Available under License Creative Commons Attribution.
Download (838kB)

Abstract

MultiLog is a logging tool that controls, gathers, and combines the output, on-the-fly, from existing research and commercial logging applications or “loggers.” Loggers record a specific set of user actions on a computing device, helping researchers to better understand environments or interactions, guiding the design of new or improved interfaces and applications. MultiLog reduces researchers’ required implementation effort by simplifying the set-up of multiple loggers and seamlessly combining their output. This in turn increases the availability of logging systems to non-technical experimenters for both short-term and longitudinal observation studies. MultiLog supports two operating modes: “researcher mode” where experimenters configure multiple logging systems, and “deployment mode” where the system is deployed to user-study participants’ systems. Researcher mode allows researchers to install, configure log filtering and obfuscation, observe near real-time event streams, and save configuration files ready for deployment. Deployment mode simplifies data collection from multiple loggers by running in the system tray at user log-in, starting loggers, combining their output, and securely uploading the data to a web-server. It also supports real-time browsing of log data, pausing of logging, and removal of log lines. Performance evaluations show that MultiLog does not adversely affect system performance, even when simultaneously running several logging systems. Initial studies show the system runs reliably over a period of 10 weeks.