Got bugs? New project lets real computer users gang up on software bugs

June 05, 2006

MADISON - Ben Liblit offers a bold prediction regarding all of the complicated software programs churning away in your computer:

They have bugs. All of them. Guaranteed.

"Software bugs are part of the mathematical fabric of the universe," says Liblit, a University of Wisconsin-Madison computer scientist. "It is impossible with a capital 'I' to detect or anticipate all bugs."

The staggering complexity of software is only part of the issue, he explains. Software has so many different points of interaction - with hardware, with networks, with other software and mostly with humans - that opportunities for buggy behavior abound.

"That behavior is so dynamic that it becomes useful to look at [software programs] almost like they were some sort of organic system, whose complete behavior is unknowable to you," says Liblit. "But there are behavior trends you can observe in a statistically significant way."

Liblit takes that metaphor literally in his research, known as the "Cooperative Bug Isolation Project." Combining programming languages and software engineering with a dose of machine learning, Liblit has enlisted real software users to tell him where the bugs are. And he has a growing "kill board" as evidence the idea works.

Liblit has created lightweight instrumentation that is added into the binary language of software. The instrumentation creates a sparse, but statistically fair, random sample of software behavior, while also protecting user privacy. The system produces regular "feedback reports" across the thousands of software programs that are in use. All of those reports get fed into a powerful database that accumulates the data and starts to identify trends.

Then, through statistical modeling techniques, Liblit is able to pinpoint software bugs that are occurring with enough regularity to affect many users. The final step in the feedback loop is a bug report, prepared by Liblit and sent back to the software engineers capable of acting on the results.

The science behind this real-world debugging approach received national attention this year. The Association for Computing Machinery (ACM) named Liblit's doctoral dissertation on cooperative bug isolation, completed at the University of California-Berkeley, as the best in the world in 2005 among dissertations nominated from both computer science and engineering programs. And while the open-source software community has begun to adopt the program, Liblit also has attracted interest from IBM and Microsoft.

The real excitement of the project, Liblit says, is that it could dramatically improve the ability to enhance software post-deployment. "Software developers deploy their programs and rarely hear directly from users, but the poor guy in tech support gets an earful," he says. "That's the only kind of feedback you get; you lob it over the wall and hope it works." This system provides direct information about real-world software problems that developers can act on with statistical certainty. Right now, the only way to gauge real-world performance is based on the "squeaky wheel" effect of those who file bug reports or call tech support. And real-world performance will always be an important variable in dynamic debugging techniques.

"It has been cynically observed that software developers use their consumers as beta testers," he says. "I think there's a lot of truth to that observation. The problem is consumers are not very good beta testers. They're not very disciplined, they don't keep good records, they never do the same thing twice. My solution is to make them better beta testers."

Right now, Liblit has a number of users in the open-source software environment, where the application has been added to popular programs such as Evolution (similar to Microsoft Outlook), Gnumeric (a type of spreadsheet), Rhythmbox (similar to iTunes) and Spin, a CPU simulator in heavy university use. He has posted 192 versions of eight different open-source applications during the three-year life of the project.

The system averages just under 3,000 new reports per month, and bug rates vary a great deal across applications, he says, with "crash rates" (where the program shuts down) as high as 8 percent in one application to a low of 0.4 percent in another. All in all, his "kill board" has recorded 546 outright program crashes and 11,369 lower-level errors as of April 2006.

Liblit says user privacy is absolutely essential to the program's future viability. The randomly sampled snippets of data have no identifiable information unto themselves, and only have meaning when aggregated in large numbers.

Will software development ever reach a level of sophistication that would render Liblit's bug machine obsolete? Liblit says that scenario is unlikely. Given that software becomes vastly more complicated with each generation, his better guess is that the industry is maintaining an even keel.

"We're trusting our software more than ever before," Liblit notes. "We're also hating our software more than ever before. And when the software fails, it's more damaging than it ever was before. But with the right technology, the users themselves can help make software better for everyone."
-end-
- Brian Mattmiller, (608) 262-0930, bsmattmi@wisc.edu

PHOTO EDITORS: High-resolution images are available for downloading at: http://www.news.wisc.edu/newsphotos/liblit.html

University of Wisconsin-Madison

Related Behavior Articles from Brightsurf:

Variety in the migratory behavior of blackcaps
The birds have variable migration strategies.

Fishing for a theory of emergent behavior
Researchers at the University of Tsukuba quantified the collective action of small schools of fish using information theory.

How synaptic changes translate to behavior changes
Learning changes behavior by altering many connections between brain cells in a variety of ways all at the same time, according to a study of sea slugs recently published in JNeurosci.

I won't have what he's having: The brain and socially motivated behavior
Monkeys devalue rewards when they anticipate that another monkey will get them instead.

Unlocking animal behavior through motion
Using physics to study different types of animal motion, such as burrowing worms or flying flocks, can reveal how animals behave in different settings.

AI to help monitor behavior
Algorithms based on artificial intelligence do better at supporting educational and clinical decision-making, according to a new study.

Increasing opportunities for sustainable behavior
To mitigate climate change and safeguard ecosystems, we need to make drastic changes in our consumption and transport behaviors.

Predicting a protein's behavior from its appearance
Researchers at EPFL have developed a new way to predict a protein's interactions with other proteins and biomolecules, and its biochemical activity, merely by observing its surface.

Spirituality affects the behavior of mortgagers
According to Olga Miroshnichenko, a Sc.D in Economics, and a Professor at the Department of Economics and Finance, Tyumen State University, morals affect the thinking of mortgage payers and help them avoid past due payments.

Asking if behavior can be changed on climate crisis
One of the more complex problems facing social psychologists today is whether any intervention can move people to change their behavior about climate change and protecting the environment for the sake of future generations.

Read More: Behavior News and Behavior Current Events
Brightsurf.com is a participant in the Amazon Services LLC Associates Program, an affiliate advertising program designed to provide a means for sites to earn advertising fees by advertising and linking to Amazon.com.