Saturday, May 19, 2007

The Real Cost of Software Unreliability

According to the National Center for Statistics and Analysis, in 2005, over 43,000 people were killed in traffic accidents in the U.S. alone. I don't know what the number is for the entire world but it must be in the six digits. No one can fault software unreliability for those fatalities since human drivers were at fault, but what if I told you that the reason that human beings are driving cars and trucks on the road and killing themselves in the process is that unreliability imposes an upper limit on the complexity of software systems? As I wrote in a previous article, we could conceivably be riding in self-driving vehicles right now but concerns over safety and reliability will not allow it. In addition, the cost of developing safety-critical software rises exponentially with the level of complexity. The reason is that complex software is much harder to test and debug.

What will it take to convince the computer industry to change over to a new paradigm that will make it possible to automate all vehicles? What will it take to convince software developers that complexity no longer has to be an enemy but can and should be a trusted friend? What will it take to convince them that there is a way to build bug-free software of arbitrary complexity? What will it take? Are 43,000 dead men, women and children not enough?

In my opinion, most of the funds allocated for traffic research by the U.S. Department of Transportation should be used to find a solution to the software reliability crisis. Why? Because the solution would keep tens of thousands of human beings from dying needlessly every year. Are you listening, Secretary Mary E. Peters?

No comments: