Poll What is the worst software bug ever?

[TMac3000]

I'm not sure of the year, but there was the missing hyphen in a FORTRAN program that caused the loss of a U.S. rocket. I think it was back in the 60's.

 

[lennartsmit]

The patch for EVE Online which replaced the boot.ini of Windows instead of EVE on some versions of XP.

Off-topic: Moach, you might want to check out this game in development: http://www.moddb.com/games/thrive

 

[Wishbone]

Reported by Siberian Tiger in another thread:
http://www.rian.ru/science/20101206/305091357.html

10:59 06/12/2010

Moscow, Dec 6 - RIA Novosti. Proton-M launch vehicle with Glonass-M satellites diverted from the target trajectory by 8 degrees because of errors in flight controlling software, programmed into the onboard computer, said us an informed source in Russian rocket industry this Monday.

"According to the very first raw information we could get, there was no mechanical problems with the Proton itself. Some specialists believe that the off-nominal regime of engines that pushed the rocket excessively and sent it off-course was caused by errors in the software programmed into the onboard flight-controlling computer." - said us the person.

He also informed us that development of mathematical and logistical support for this launch was carried out by specialists of Pilyugin Scientific Development and Production Centre for Automatics and Instrumentation (NPTs AP). {...}

Reported by orb:

http://www.interfax.com/newsinf.asp?pg=6&id=207016

MOSCOW. Dec 6 (Interfax-AVN) - The Moscow-based Pilyugin Automation and Instrumentation Production Center has denied reports that failures in the control system caused the accident with the Proton-M launch vehicle carrying three Glonass-M satellites.

"We have already checked that the control system and the software operated regularly. The accident must have been caused by some outside factors," said a spokesman for the center that developed the control systems of the Proton-M LV and the DM upper stage used for launching Glonass-M satellites.

That was his comment on media reports that a failure in the control system and control software caused the satellite accident.

Among outside factors he named "low-quality fuel or erroneous data concerning the weight of the head part of the rocket."

 

[Linguofreak]

An interesting case is that of the Therac 25.

 

[Frogisis]

Any bug that happens just when I'm trying to save my work is the worst bug.

 

[n72.75]

What was the highest loss of life due to a software bug?

What was the highest cost due to a software bug?

 

[Wishbone]

Thanks to Linguofreak for bringing it up. Therac 25 and the Sayano-Shushenskaya catastrophe are probably the ones with the highest toll.

 

[Nazban]

For me the biggest bug is with the online EA store(sims 3)
we pay money to bug stuff from their store that cannot be downloaded.:compbash2:
Many people have complained to their customer service with the most silliest answers(like if my internet connection was on). And this has been going on for almost a year...

 

[Urwumpe]

What was the highest loss of life due to a software bug?

Should be the 75 deaths of this one:

https://secure.wikimedia.org/wikipedia/en/wiki/2009_Sayano-Shushenskaya_hydro_accident

Alternatively, you can also count many Airbus A320 problems as software bugs, often it had been related to the non-standard interface or poorly documented behavior of the flight controls

What was the highest cost due to a software bug?

Likely something trivial and silent, like problems in SAP or electronic tax reports.

 

[Sky Captain]

How is Sayano Shushenka accident a software bug? IIRC the cause was faulty turbine and metal fatigue.

 

[Urwumpe]

How is Sayano Shushenka accident a software bug? IIRC the cause was faulty turbine and metal fatigue.

The final failure of the turbine was caused by the software failing to regulate the turbine pressure in controlled regions, a small bug in the operating software did not close the inlet valve, the turbine was literally overspeeding into a region, that was not recommended for operation, even for a intact turbine.

On the day of accident the turbines worked at working level 212 metres (696 ft). At this pressure the recommended power band for the turbines is 570–640 MW (band III) and the allowed band is 0-265 MW also (band I). Band 265–570 MW (band II) at this pressure is not recommended and output over 640 MW (band IV) is forbidden. On the day of accident turbine 2 worked as the plants power output regulator and due to this, its output power changed constantly. The turbine often operated in band II regime which is accompanied with pulsation and strokes of water flow.

It was initially in band III, was dropped into band II (again), in which the control of the turbine failed, because of the pulsations. Quickly reducing the power output of the turbine results in the rotor being pushed upwards (on all such turbines), which is why you have the software to prevent entering such regions during normal operations.

If the turbine would have been intact, the damage would have been contained inside the reinforced turbine structure, but in this case, it was no longer contained, causing the catastrophe. Such events are always a chain of failures on multiple points, there is no single event that causes high damage.

 

[Wishbone]

Nancy G. Leveson paper to appear in AIAA JSR

The Role of Software in Spacecraft Accidents. The preprint is available here:

http://sunnyday.mit.edu/papers/jsr.pdf

My heart-felt thanks to Prof.Leveson :tiphat:

 

[SiberianTiger]

Are you involved somehow?..

 

[Wishbone]

No. That's the woman who wrote the investigation report on Therac-25 "death ray" machine.

EDIT: On further thought, and at the mods' discretion, this could be appended to the "Costliest bug" thread.