The Software Bug All Stars - and what we can learn from them

NASA lost a $125 million Mars Orbiter because one team used metric units while another used imperial. We'll dissect this and other famous bugs for their valuable lessons.

#1about 1 minute

The high cost of software bugs and why we study them

Software bugs cost the economy billions annually, but studying major failures from other companies provides valuable, and sometimes entertaining, lessons.

#2about 8 minutes

How a unit conversion error destroyed the Mars Climate Orbiter

A mismatch between imperial and metric units caused the probe's failure, highlighting the need for typed value objects and clear communication channels.

#3about 5 minutes

The fatal race condition in the Therac-25 radiotherapy machine

A race condition combined with poor error messages and a lack of independent QA led to patient deaths, teaching lessons about humility and organizational safeguards.

#4about 6 minutes

Handling date and time edge cases in software

The Microsoft Zune's leap year bug and an iOS alarm failure demonstrate the complexity of date handling and the importance of using robust libraries.

#5about 4 minutes

Preventing resource leaks in long-running systems

The Boeing 787's need for a periodic reboot due to data accumulation serves as a reminder to manage and clean up resources like log files.

#6about 2 minutes

What we can learn from high-profile software failures

Even large, experienced organizations make critical mistakes, so developers should actively learn from these public failures to improve their own work.

#7about 2 minutes

Q&A: A bug caused by exhausting the Linux randomness pool

The speaker shares a personal story about a bug where excessive requests for random numbers exhausted the system's entropy pool, causing the application to block.