The TaskTracer project at Oregon State University is part of the Management of Knowledge-Intensive Dynamic Systems (MKIDS) initiative funded jointly by the National Science Foundation and the Intelligence Community (a coalition of federal intelligence agencies). We are investigating the possibilities of a desktop software system that will track in detail how knowledge workers complete tasks, and intelligently leverage that information to increase efficiency and productivity.
Our goal is to develop five capabilities:
- More task-aware user interfaces in the applications we use daily
- More efficient task-interruption recovery
- Better personal information management
- Workgroup information management
- Within-workgroup workflow detection and analysis
There are a substantial set of research challenges that must be faced in order to successfully develop tools with these capabilities. These challenges include user interface design, machine learning, privacy and workplace culture, data collection, systems architecture, and data modeling. Through the development, testing, and deployment of TaskTracer in real workplaces, we expect to contribute substantially to knowledge regarding intelligent personal desktop information systems and the underlying technologies used.
User Needs
Knowledge workers spend the majority of their working hours processing and manipulating information and are highly multi-tasking. The information they manipulate may be encoded in many different formats: documents, software code, web pages, email messages, phone conversations. At the center of the TaskTracer project is the concept that almost all knowledge workers organize their work into discrete and describable units, such as projects, tasks or to-do items. Our approach will combine user input, creative user interfaces, and machine learning to assign each observed action (opening a file, saving a file, sending an email, cutting and pasting information, etc.) to a task for which it is likely being performed. Once we have the past events structured by task, we can provide substantial value to the knowledge worker in assisting in their daily task routines.
Making Interfaces Task-Aware
Current desktop applications and tools were not built to support multi-tasking workers. They assume that the set of tools and documents that you need to access remains consistent. However, when users are highly multitasking, the set of resources that they need may change rapidly – every time the user’s current task changes. In TaskTracer we envision a desktop environment that is task-aware, where applications known that users have tasks, each with different needs, and are aware of when task switches happen. If an application is task-aware, it will know what files you are likely to open, what directory you are likely to need to save a file in, what people you may write emails to, and so on. Support for such task-aware activity could greatly boost productivity.
Interruption Recovery and Personal Information Management
By definition, multitasking people face continual interruptions as they switch between ongoing tasks. Given that knowledge workers are involved in nontrivial analysis and deal with large amounts of information, recovering from interruptions often has a significant overhead cost. This overhead may be cognitive: workers may have to remember exactly where they were in a chain of logic, or why they decided to take their most recent action on a task. The overhead may also just lie in the manual interaction needed to locate and access the necessary resources (e.g., documents and/or software tools).
In terms of interruption recovery, when a user resumes a task, TaskTracer can help users regain the cognitive context they held prior to switching away from that task. In essence, TaskTracer has the ability to “replay” significant actions (and their results) that the user performed immediately prior to being interrupted.
Because TaskTracer has complete records of how the user performed past tasks, these past tasks can serve as a valuable personal knowledge repository. In addition to just serving as a searchable, structured record of past actions, the TaskTracer can make use of its knowledge of past task sequences as templates for the user’s current work. This can be highly effective because knowledge workers frequently repeat similar task sequences over and over and frequently make use of past documents as templates for new documents. We seek to design software that helps people to rapidly locate, discover, and reuse past processes they used to successfully complete tasks.
Workgroup Knowledge Management and Workflow Detection
When a workgroup of knowledge workers interconnect their personal TaskTracers, several additional capabilities can be explored. First, each member of the workgroup can now leverage not only their own task records, but also the task records of all other members of the workgroup. This allows detailed information about work processes to be conveniently shared between users. Once a history of tasks records across the workgroup have been collected, the data in the task records can be used to explore and model workflows between members of the workgroup.
User studies
We have built an initial TaskTracer prototype for the Microsoft Windows environment. The TaskTracer prototype hooks into the popular Microsoft Office tools, as well as Internet Explorer and Windows Explorer. Members of the research team have been running TaskTracer for several months now, and most are now heavily addicted. We have several user studies in progress or in planning. We are currently involved in a semi-structured interview study of knowledge workers to establish an understanding of current user perceptions and expectations of task and information management. We are preparing to perform an experiment to understand the cost (in terms of low-level user actions, as well as more high-level costs such as cognitive operations) of activities that can be avoided with TaskTracer. We are also planning to deploy TaskTracer to production desktops of knowledge workers within Intel, collecting data to understand not only how current knowledge workers multi-task, but also understand the potential of the TaskTracer tool.