Margaret M. Burnett, Research

Recent and Current Projects

• End Users and Machine Learning. We have been investigating how to support end users of personalized systems involving machine learning to understand why these systems make the decisions they make, to assess whether and when to trust these systems, and to fix problems they see in these systems.

  For more on this, see my papers page. Anything since 2007 co-authored with Simone Stumpf, Weng-Keen Wong, or Todd Kulesza is about this work.

• Information foraging theory: In recent years, the software engineering community has begun to study program navigation and tools to support it. Some of these navigation tools are very useful, but they lack a theoretical basis that could reduce the need for ad hoc tool building approaches by explaining what is fundamentally necessary to the people using such tools. We've been working on applying information foraging theory to this problem in the domain of debugging. Toward this end, we've developed an application of the theory to debugging navigation, and have built a predictive model we call PFIS (Programmer Flow by Information Scent), a high-level behavioral model and algorithm of programmer navigation during software maintenance. We're currently doing a lot of empirical work on expert programmers debugging real bugs to evaluate the effectiveness of this theory and the PFIS model.

  For more on this, see my papers page. Anything since 2007 co-authored with either David Piorkowski or Joseph Lawrance is about this work.

• Gender HCI. What if problem-solving software features are designed in such a way as to be more conducive to one gender's problem solving styles than the other's? We are working to understand whether there are some aspects of software features that interact with gender differences, and if so, how to change these software features so that both genders are well supported. See the web page for our results. You can also go to my papers page; anything co-authored with Laura Beckwith or Valentina Grigoreanu is about this topic.

• End User Software Engineering. My collaborators and I were the first to take a serious look at software engineering aspects of end-user programming. In 2003, we joined together with other researchers to form the EUSES Consortium (End Users Shaping Effective Software), which is a group of researchers from Oregon State University, Cambridge University (UK), Carnegie Mellon University, City University London, Drexel University, Penn State University, University of Nebraska, University of Washington, National Instruments, and IBM, working together on this problem. The research question: Is it possible to bring some of the benefits of rigorous software engineering methodologies to ordinary end users?

  For more on this, see the EUSES Consortium site.

Here's some of my past research:

• WYSIWYT. With colleagues Gregg Rothermel, Curt Cook, and Thomas Green, we spent several years working to bring at least some of the benefits of formal testing to an audience of end users and programmers who are not trained in formal testing. "WYSIWYT" stands for "What You See Is What You Test". See the WYSIWYT web page for information about it, including the papers and patents we have on it.
• I worked with Jon Herlocker and Tom Dietterich on the TaskTracer Project. TaskTracer aims to help organize the user's work by task to achieve benefits such as easing re-entry into a task worked on before, recovery from interruptions, reuse of information about a task.
• I spent a number of years doing visual language design and end-user programming language design,
  • mostly in the context of my visual spreadsheet language Forms/3.
  • My students and I also developed another visual spreadsheet langauge, this one for web applications, called FAR.
  • I was also part of a spreadsheet research project with a redesign of how users might design their own functions in Excel. Here is the first paper on that, and here is the second.
  • Some of my end-user software engineering work (and that of my colleagues, Martin Erwig and Gregg Rothermel) ended up in a commercial spin-off company, RedRover Software, whose products aim at helping end users with the dependability of their spreadsheets.
  • I've also done a couple of studies on the CoScripter web application builder for end users.
  We still use Forms/3 for some of our end-user software engineering work.

Visual Language Information:

A paper introducing visual programming.

Publications:

Most of my publications are listed on my publications page, and many are also listed in the visual language bibliography.

Acknowledgments

Date of last update: Dec. 27, 2011.