Posts

Now in its third decade, the Association for Computing Machinery (ACM) Conference on International Computing Education Research (ICER) is widely considered the leading conference in its area. Last week, the organizers of ICER 2026 awarded work from 2014 (“The Recurring Rainfall Problem”) by Brown CS faculty member and Director of Undergraduate Studies Kathi Fisler with their Lasting Impact Award, noting that it demonstrates how programming languages impact how frequently students make mistakes. The award recognizes a specific ICER research paper published at least ten years prior that has made a sustained, significant, and meaningful impact on the field of computing education research.

In the solo paper, written when Kathi was a faculty member at Worcester Polytechnic Institute, she returns to and makes progress on Soloway’s Rainfall problem, often used to assess student programming skills in introductory courses. Soloway’s original read, “Write a program that will read in integers and output their average. Stop reading when the value 99999 is input.” 

“Rainfall is interesting,” Kathi says, “because it's conceptually straightforward, yet with enough components to raise subtleties in code. Historically, students have performed poorly on this problem.” During a research visit around 2012, leading CSEd researcher Mark Guzdial asked Kathi whether anyone had tried Rainfall with students trained in the How to Design Programs curriculum (co-authored by Brown CS faculty member Shriram Krishnamurthi, and which Kathi had taught for many years). Within a couple of weeks, Kathi had set up the study in five courses across four universities.

“Back in 2014,” Kathi says, “few CSEd studies had explored students who were learning functional programming; this style has built-in operations on lists (such as maps and filters) that suggest ways to decompose programming problems. I proposed a novel two-level analysis method for extracting code structure (plans) and implementation details.  My participants produced solutions with rather different structures than in earlier studies, and seemed to make fewer low-level errors. At first, people asked whether my findings suggested that ‘functional programming is better’, but the followup study we did, using multiple planning problems and different groups of students who were learning one of Java or functional programming, showed that the essential variable was actually the set of built-in and library operations that students knew for processing lists and strings. In the original Rainfall studies, students had only arrays and keyboard I/O, so they had to express the program at a lower-level of granularity. Naturally, that is more error-prone.”

After Kathi’s 2014 paper, several other groups went on to conduct Rainfall and planning studies, thus reviving a 30-year-old area in modern coding contexts.

Kathi’s interests center on the interaction between human learning and formal structure, drawing on a combination of cognitive science, education, formal methods, data structures, and socio-technical systems. Her attention is currently on how to integrate the power of AI coding agents into early CS courses while still helping students develop the conceptual structures needed to reason about programs on their own. She has an almost thirty-year history of outreach for K-12 computing education, mainly through Bootstrap and various standards committees, and she’s one of the lead authors of A Data-Centric Introduction to Computing. Her recent accomplishments include being named an ACM Distinguished Member, receiving the inaugural SIGPLAN Distinguished Educator Award (2023, joint with Shriram Krishnamurthi) and joining the Editorial Board of Communications of the ACM to handle Education-related articles.

For more information, please click the link that follows to contact Brown CS Communications Manager Jesse C. Polhemus.