School of Electrical Engineering & Computer Science

Research Areas
End-User Programming, Software Engineering, Visual Programming Languages

Research Description
By 2012, over 90 million people will use computers in American workplaces, including at least 55 million who will create spreadsheets, databases, scripts, or other programs to automate tasks that would be tedious or error-prone to do manually. For example, an end-user programmer might create a script that computes the monthly interest due on a loan.

However, end-user programmers run into trouble when trying to automate larger tasks. In particular, combining several programs into a larger program is often impossible or requires intricate, brittle code. For instance, existing programming tools offer no easy way to use borrower names in a spreadsheet to look up corresponding loan information from an existing database provided by a co-worker, to compute each borrower’s monthly interest payments by repurposing the script mentioned above, and to display the results beside each borrower name in the spreadsheet.

My goal over the next few years is to enable end-user programmers to create, share and combine code. Two key stumbling blocks stand in the way. First, end-user programmers need a reliable way of finding existing code that they can adapt and combine. I will to extend a preliminary model for identifying reusable code, then use this model to provide methods and systems for guiding end-user programmers to locate high-quality code. Second, after identifying useful existing code, end-user programmers need a reliable way to integrate code. I will focus first on helping programmers pass data from one piece of code to another. As an initial step, I developed a data model and supporting tools in my thesis work so that end users can create code for validating and reformatting the many string-like data values in end-user programs. These strings are modeled with abstractions called “topes” that are platform-independent and which might ultimately serve as a “lingua franca” for exchanging data between scripts, spreadsheets and other kinds of code.

My thesis work also exemplifies my research approach, which is to identify a general problem by studying the needs and context of human beings, and then to solve the problem by adapting the models and methods of software engineering and human-computer interaction.