No matter how hard we rely on machines, they are far from ideal. At least because they are made by man.
Until programs become smarter than people, we will have to monitor their status constantly. Especially when we are talking about a responsible business.
The stories below will show how much our lives depend on a properly assembled algorithm, and what happens if quality control in production takes a day off.
Let's start with the strange things affecting a relatively modest number of people, and end with real financial disasters. They cost more than a dozen sleepless nights for those who had to urgently read a fragile code and literally save humanity.
1. The system buried alive 8,500 hospital patients in Michigan
In 2003, the St. Mary of Mercy Medical Center in Grand Rapids updated its patient registration program to a new version. Due to misinterpretation of the data, the variables "discharged" and "died" are confused.
Therefore, everyone who has already undergone treatment began to receive death notifications by mail and in various reports like a blood test.
Promotional video:
The problem would not have become large-scale, but due to the high automation, messages were sent to both patients and insurance services. When the latter saw that the person was "dying", they stopped compensating for subsequent treatment. This included more than 2,000 pensioners and disabled people.
Since the error was found two months after the breakdown, it took about the same amount to restore information in related services and compensate for the damage.
2. The software update has deprived 60 thousand people of long distance calls
In January 1990, the American telecom operator AT&T improved its tower switch monitoring program. Due to an error in the code, one of them during the call began to send signals faster than the other could process them.
The data began to overlap, and the problem quickly spread to other points. At the other end, people only heard noise. This went on for 9 hours.
The problem was solved by rolling back the software to the previous version, but the problem has not ceased to be relevant.
The situation repeated itself at least once in 1998, but then only SMS service notifications were affected.
3.5% of all stores in Russia broke down due to a new online checkout
On December 20, 2017, the data control technique was updated so that sellers began to directly transfer information about transactions to the tax office. One such improvement has blocked every 20 store in the country.
Failures began in the salons of the DNS network in Vladivostok, where people wake up before Moscow. The system did not allow sending payments to the Federal Tax Service (FTS), and because of this, cashiers were not allowed to sell goods.
While the problem reached the capital, from where the issue began to be resolved, some points of Magnet, Pyaterochka with Perekrestok, Eldorado and Rigla pharmacies arose throughout Russia.
The Federal Tax Service had to react quickly and allow stores to work offline. Those were allowed to enter data after the system was restored.
The problem was completely eliminated within a few days with patches and forced reboots.
The theoretical damage, according to the Association of Internet Trade Companies, could reach 2.5 billion rubles. The real one turned out to be slightly lower due to the rapid optimization of processes by the Federal Tax Service.
4. The car was given to design a stadium in Connecticut. He collapsed
Since 1972, the City of Hartford has tried to expand its infrastructure and invested in major projects. One of them is the Hartford Civic Center - a complex of shopping, entertainment and sports grounds.
The structure of the stadium was designed through the program, which, together with the optimized consumption of materials, saved the city about $ 500 thousand.
The complex has been fully operational and has even been the "home" of the local hockey group New England Whalers since 1975.
However, on the morning of January 18, 1978, the stadium collapsed. There were no games that day: the building was empty and no one was hurt.
The media spread the news, where the reason was the severity of the snow. But the investigation showed that in fact the problem was complex, and the root was naive trust in the program.
The four supporting columns have been poorly thought out in size and support since their construction. The stadium began to gradually "take shape" even during construction, and quality control teams were distributed among different contractors and poorly coordinated data.
The restoration cost the city $ 90 million. Subsequently, the XL Center arena was erected on the site of the complex, which still serves as the main sports ground in Hartford.
5. Intel released a buggy processor and caused an international scandal
In 1994, a Pentium-branded CPU was the company's flagship, and it hid a microscopic problem that affected a tiny fraction of people: when a user divided one number by another, the result was wrong. The error looked like this:
The programmers misconfigured one of the branches of operations hardwired into the processor. She looked for root data and found the wrong ones.
In this case, the main damage fell not on users, but on the company.
Due to the fact that Intel already then felt confident in the market, and the chips were new, even the federal media in many countries picked up the news and caused catastrophic damage to the company's image and income.
As a result, in 1994, the replacement of all damaged processors reduced the company's revenue by half of the planned amount - by $ 475 million.
6.6 million cars may not deploy airbags
In January 2020, it turned out that sensors in some Toyota and Honda models were too sensitive to electrical noise.
There is a possibility that at the moment of a collision the system will not signal the security system. She will not be able to keep the belts taut, and the pillows will not be filled with air.
The problem could be more global, since the computer from Toyota cars was developed by a third-party organization ZF-TRW. And she supplied her developments to at least six companies in the United States alone, which sold 12.3 million cars.
But so far only Japanese manufacturers have decided to repair the sensors. And then, many are still waiting for a notification from their dealers.
7. MySpace destroyed 50 million user songs
In 2016, the company was doing data migration, which began back in 2013. Even then, some materials and accounts became inaccessible to some users.
And during the transfer of his huge library of music, photos and videos, an error occurred on the server, which permanently deleted them.
Since the management did not disclose the exact reason, there are various rumors about the source of the problem. For example, there is an assumption that keeping such a volume of old data is unprofitable and it was cheaper for management to “accidentally” delete it.
One way or another, the world lost one of the largest layers of internet culture from 2003 to 2015.
8.14 thousand single parents did not receive government payments
In April 2003, Child Support Agency, a UK-based company for low-income and disadvantaged families, introduced a system to filter applications. It was worth £ 300 million.
Six months later, it turned out that less than one-twentieth of the requests were processed, and many children were left without material assistance.
The scandal lasted until at least 2006, when the program continued to eat up 70% of the money allocated to the project and the costs by 2010 were £ 1.1 billion.
As a result, in 2012, the agency was closed and a new organization, Child Maintenance Group, was launched in its place.
9. Vulnerability in the protection of 500 thousand largest sites gave access to your RAM
In April 2014, security experts discovered a critical security hole in the OpenSSL library, which powers the most widely used HTTPS protocol.
It was named Heartbleed after the Heartbeat process used as the basis for this error.
Using the vulnerability, it was possible to find out what is in the victim's computer RAM.
And, although the maximum amount of stolen information could not exceed 64 KB per request, this was enough for access to passwords and confidential messages.
The bug affected 17% of all protected sites. Including Google, Facebook, Instagram, Twitter and even Minecraft.
The danger was covered with a simple patch, so many companies reacted quickly.
However, only one problem is comparable in scale to this problem, and you have probably heard about it at least once.
10. The world spent $ 300 billion to keep computers working in 2000
Until 1999, systems were programmed so that some marked dates in an 8-digit format (HH. MM. YYYY), while others left 6.
Photo by Emory Kristof / Emory Kristof.
This could lead to the fact that the transition to the new millennium would cause errors in programs across the planet.
The date of the format HH. MM. YY could replace 2000 by 1900, since both numbers end in "OO". Thus, the error would rewrite and erase the data, disrupt algorithms and provoke the collapse of online systems.
Most of the companies' time and resources were spent not on fixing the consequences, but on checking every computer in the company.
Because software hadn't experienced such leaps in time before, the situation was discussed around the world.
There has been a lot of talk around the Year 2000 (or Y2K) Problem, including the advisability of panic. They were fueled by the fact that countries took the issue seriously and prescribed initiatives at the state level.
For example, Russia has created an official document, the National Action Plan to Address the 2000 Problem in the Russian Federation.
The scoreboard on the last line is "zeroed" and shows 1900 instead of 2000 /
The nearest similar error will overtake non-optimized 32-bit systems in January 2038, but programmers are already preparing for the transition.
The situation will affect 64-bit systems in 292 billion years, so you can relax here.
Much more real and rather threatens the Problem of 10,000 with its transition to five-digit values. It seems that there is no need to worry about it - for now, the question is rather theoretical.
True, there is always the possibility that bits of existing code from your smartphone will survive to that era.
Maybe it's worth thinking about it.
Author: Pavel Teleshevsky