The Worst Software Bugs In History

[filkertom]

Here.

Comments

[shockwave77598.livejournal.com]

They overlooked my favorite one. Not surprising, actually, since the goof destroyed a 1.4B$ spy satellite.

[partiallyclips.livejournal.com]

I'm surprised that the crashes of several commercial aircraft, particularly that Airbus debut at the Paris air show a decade or so ago, were not mentioned. Some of them are here:

http://www.baselinemag.com/article2/0,1540,1543590,00.asp

Not to mention the US Navy shooting down that Iranian Airbus, which was at least BLAMED on a software bug.

[phillip2637.livejournal.com]

Try http://groups.google.com/group/comp.risks (also known as ACM forum on risks to the public in computers and related systems). There are years of horror stories of the "when will we ever learn" variety.

[shockwave77598.livejournal.com]

In the late 90s, there was a team management approach called compartmentalization. In this, each team had a tiny section of code or hardware to deliver. And as long as everything used the APIs correctly, the theory went, the program would work perfectly and nobody would be able to see everything at once. This held much appeal with makers of spy things like Lockheed.

Unfortunately there's a down side to not having anyone able to see the whole thing as a system. Code was taken from one flight vehicle and was reused in the latest vehicle. No problem there. And all the APIs worked correctly and the code was essentially bulletproof. Again, no problems.

But the system the code was taken from had a hydraulic pump, whereas the new vehicle had only an accumulator. And the software kept making the thousands of engine adjustments as it always had, not knowing that it had a finite oil supply and needed to use it very sparingly.

Finally, it ran out of fluid, the nozzles turned and locked, and the vehicle went out of control and was destroyed. Price tag - 1.4Billion dollars lost. All because some manager thought software was put together like tinkertoys.

I don't know of any glitch costing more than this one.

[kengr]

*Technically* speaking that was a failure in writing the specs for the code. They should have explicitly stated that the limit existed.

Then again, a lot of bugs are failures in the specs. And most of ther rest are failure to *conform* to the specs.

Sony Music CDs now install trojan on computers

[rothesis.livejournal.com]

OK, so strictly speaking, this isn't a bug. However, I was surprised you had not mentioned this on your Journal.

Since early 2005, Sony music CDs have included a digital rights management package that installs itself as a "rootkit" on any computer playing the CD.

Here are the links where I got my information:
http://www.langa.com/newsletters/2005/2005-11-07.htm#7
http://www.theregister.co.uk/2005/11/01/sony_rootkit_drm/

As near as I can tell, Sony & the RIAA want to ensure that filk is the only music I buy.

I have seen that bit....

[filkertom.livejournal.com]

Ohhhh, that's not a "bug". That's an "undocumented feature". And it's just another example of the music industry in general, and Sony in particular, wasting money and pissing off their customer base. And, just like the last effort (remember, the one that could be circumvented with a Sharpie?), this one isn't gonna be around long, either. Hackers have too much fun defeating copy protection schemes.