A multi-university consortium led by Princeton University computer scientist Andrew Appel aspires to stamp out software bugs with the help of a five-year, $10-million grant from the U.S. National Science Foundation.
The goal of the Expeditions in Computing: The Science of Deep Specification (DeepSec) project is to develop integrated tools to eliminate uncertainty from software development, with implications for reshaping industry by bridging the chasm between researchers and educators.
The team will initially focus on learning what factors determine how various computer elements interoperate. They then will devise "deep specifications" to enable engineers to build bugless programs as well as confirm they behave as intended.
Appel says his work in an earlier project yielded a program that can accurately translate programming language into machine instructions, which are executable on a computer chip. "The logical next step is to connect verified components--compilers, operating systems, program analysis tools, processor architectures--so no bugs can creep in because of misunderstandings at component boundaries," he says.
DeepSec will accomplish this by improving how specifications are written using formal logic. "But to test whether this approach is really an improvement, we need a big consortium with multiple components to connect together," Appel says.
From Princeton University
View Full Article
Abstracts Copyright © 2016 Information Inc., Bethesda, Maryland, USA
No entries found