Abstract: The use of advanced computational tools in every discipline is exploding. Scientists are building sophisticated programs to analyze and predict biological phenomenon, to learn climate trends, and to understand chemical interactions. Businesses are utilizing machine learning libraries to create algorithms that both predict client behavior as well as enhance their customers’ experience. Hence, we are looking to a future where everyone is a programmer and programming is everywhere. However, software remains inherently complex, and this new breed of software suffers from weakly known oracles, heavy reliance on data, a trust in a closed box view of computation, and excessive configurability. Software is likely to be buggy, unreliable, non-reusable, flaky, and frankly – surprising. Rather than lead a new generation of end-user-developers to the edge of a chasm with few safety ropes in tow, we have an opportunity now to design for the future. We should be designing novel ways to pry open the magic boxes, creating non-intrusive techniques for explainability and interpretability, provide easy-to-deploy validation techniques that empower domain experts, and build in the easy creation and inclusion of meta-data, documentation and first class configurability.

Bio: Myra Cohen is a Professor in the Department of Computer Science at Iowa State University, and the Lanh & Oanh Nguyen Endowed Chair of Software Engineering. Her research interests are in software testing of highly-configurable software, search based software engineering, and synergies between systems and synthetic biology. She was recently selected as a Better Scientific Software Fellow to help improve software testing in scientific domains. She an ACM Distinguished Scientist, and serves on the steering committees of ASE and ESEC/FSE, and is active in other software engineering conference organizational roles. She was the program co-chair for ESEC/FSE 2020 and ICST 2019 and was the general chair of ASE in 2015.