Biggest Software Failures of All Time

Software Development companies across the globe invest a huge amount or budget and resources in software testing. Despite all the efforts, some of the bugs are not discovered and lay dormant in the software program. This is due to the inherent limitation of software testing, as it cannot be carried out for the large number of possible execution paths. The bug is triggered later some time once the non-tested bug triggering scenario is executed. The bugs triggered may have a huge consequence costing a large amount of money to the organization or country. Some of the software failures which had a huge impact are listed below: 

“To err is human, but to really foul things up you need a computer.”  

Paul Ehrlich 

1. MARS CLIMATE ORBITER 

The $125 million Climate Orbiter spacecraft was lost in space during its mission to Mars in 1998. The communication with spacecraft was lost because it went into orbital insertion 286 days after its launch. The navigation error pushed the rocket too close to the Mars atmosphere where it burned and broke into pieces.

Reason. It was revealed that the software of ground-based controller generated output in non-SI units of pound-seconds instead of metric units of Newton-seconds.

2. ARIANE 5 FLIGHT 501

The European heavy-lift Ariane 5 rocket reused working software from its predecessor, the Ariane 4. The rocket was unable to achieve orbit and exploded few seconds after launch. The Ariane 5 had cost nearly $8 billion to develop, and was carrying a $500 million satellite payload. Thirty-six seconds into its maiden launch the rocket’s engineers hit the self destruct button following multiple computer failures. 

Reason. The whole system crashed when a software program tried to convert the sideways rocket velocity from 64-bit to 16-bit format. The number was too big, causing an integer overflow. When the first computer failed, control was passed to an identical redundant unit, which also returned a failure because it was running the same software. As a result, flight got diverted from its original path 37 seconds after launch. 

3. THE MARINER 1 SPACECRAFT 

The  $19 million Mariner 1 spacecraft was launched towards Venus in 1962, which barely made it out of Cape Canaveral and diverted from its intended flight path. NASA ground controllers issued a self-destruct command as it was crashing back to earth.

Reason. Investigation board later determined that the omission of a superscript bar in software program allowed the transmission of erroneous guidance signals to the spacecraft. A programmer incorrectly programmed a handwritten formula, missing a single superscript bar.  Without the smoothing function indicated by the bar, the software treated normal variations of velocity as if they were serious, causing faulty corrections that sent the rocket off course.

4. PATRIOT MISSILE ERROR

In February of 1991, a U.S. Patriot missile defense system in Dharan, Saudi Arabia, failed to intercept and track an incoming Iraqi Scud missile on Army barracks. The attack killed 28 American soldiers.

Reason. A government investigation report determined that a software problem resulted in an erroneous tracking calculation. On the day of the incident, the inaccuracy was serious enough to cause the system to look in the wrong place for the incoming missile. Prior to the incident, officials had fixed the software to improve the Patriot system accuracy. That modified software reached the base the day after the attack.

5. PENTIUM LONG DIVISION BUG

A math professor discovered a flaw in the floating point unit of Intel’s Pentium processor in 1994. For precise floating point calculation, the processor would return erroneous decimal values. There were around 5 million faulty units in circulation and Intel decided to replace all chips for anyone who could prove they were affected. Later, Intel turned some of their faulty processors into key chains. The episode cost Intel $475 million along with lost reputation.

Reason. The divider in the floating point unit of Pentium processor had a flaw in division table, missing five entries out of around a thousand. Error was only likely to occur once in nine billion random floating point divides. Intel calculated that the error caused by the flaw would happen so rarely that the vast majority of users wouldn’t notice.

6. BRITISH PASSPORT SYSTEM

The United Kingdom passport agency started a new computer system in 1999, which failed to issue passports on time to more than half million citizens. Later, the agency had to pay almost $14 million in various forms to the people.

Reason. The agency released their software in a hurry without testing and without training the employees on how to use the new system. The new law (during that time) required all children under 16 travelling abroad to get a passport, which resulted in a huge passport demand, over burdening the new software system.

7. Y2K BUG

The Y2K bug, also called Millennium Bug, was a problem in the software of computers and computer networks across the world at the turnover of calendar from year 1999 to 2000, causing a damage of around $500 billion. 

Reason. Many legacy software programs used two digit numbers to store the year for dates, for example “98” for 1998. This caused the programs to operate incorrectly after 31st December 1999.

8. RADIATION THERAPY DEVICE FAILURE

The Therac-25 medical radiation therapy device malfunctioned and was involved in many cases where hundreds of patients were exposed to overdoses of radiation between 1985-87. This failure resulted in a loss of human life and injured many others.

 

Reason. The therapy planning software calculated radiation dosage based on the order in which data was entered, often delivering a double dose of radiation, due to a subtle bug called race condition.