Talk by Chris Bogart - Eliciting Informal Specifications from Scientific Modelers for Evaluation and Debugging

TITLE: Eliciting Informal Specifications from Scientific Modelers for Evaluation and Debugging

SPEAKER:  Chris Bogart, Oregon State University

HOST: Gail Murphy

LOCATION: ICICS/CS X836 

DATE AND TIME: Friday, June 21, 1:30 pm

ABSTRACT:

Scientific modelers do not have access to the arsenal of
techniques for checking their specifications that professional
software engineers have, such as unit testing and assertions. These
techniques require tools, motivation, experience and training that
programmers without professional software engineering training may not
have. As a result, modelers face greater hurdles in debugging and
validating the programs they write. In this talk I introduce the
concept of “evaluation abstractions” as a framework for tool designers
to think about this kind of support. Evaluation abstractions are the
patterns of data in program traces and outputs that programmers
examine in order to evaluate program behavior.

I will discuss a theory of evaluation abstraction support (EAST) that
describes the factors contributing to a modeler's decision to use an
evaluation abstraction support feature; and a user-centered design
methodology, Natural Programming Plus (NP+), for the design of
interactive languages aimed at experienced users in a way that allows
for validation early in the process.

I will also describe the design and evaluation of a tool for cognitive
modelers (psychologists who study human cognition by writing
simulations of cognition), that elicits and persists a database of
modelers’ evaluation abstractions, in a piecemeal, just-in-time
fashion as their questions about model behavior arise, and uses
modelers’ own evaluation abstractions to structure visualizations,
listings, and regression tests as they continue to maintain and
develop their projects.

BIO:

Chris received his computer science PhD in June from Oregon State
University this month, pursuing research in the human-computer
interaction aspects of the design of programming languages and tools
with Dr. Margaret Burnett. He has a decade of experience as a software
engineer and previously received an MS in computer science from
Colorado State University in 1992, where he studied genetic algorithms
and neural networks with Dr. Darrell Whitley. He is a student member
of the EUSES Consortium (End Users Shaping Effective Software).