Y2K

The UAE government has said 'we have succeeded in solving the millennium problem in all state offices'. Meanwhile, authorities in California, USA are busy enacting legislation protecting computer manufacturers' liability in Y2K cases. Al Shindagah investigated the phenomenon that is Y2K.

Depending on who you talk to, the year 2000 computer bug, or Y2K to give it its official term, will cause catastrophes on a global scale or simply cause minor inconveniences. The problems posed by the Y2K bug are fuelling Doomsday prophecies. Planes will fall out of the sky, trains will go off the rails, phone networks will go down and banks will assume its accounts have been inactive for 100 years and shut them.

So what is causing all the fuss? The problem surrounds computer chips' inability to handle four-figure dates. As a result, after 31-12-99 many computers will register the date as 01-01-00, and start over from 1900.

Although this seems like an unimportant detail, it could have disastrous consequences on a global scale. The US and Russia recently announced plans to deal with the computer bug in defence systems that are heavily reliant on computers to combine data from satellites and radar. This data is used to monitor impending threats such as missile launches and unidentified aircraft. And because Russia and the US still keep around 2,500 nuclear missiles available for launch at short notice it is imperative the computers that handle this information do not panic and launch an arsenal.

Unfortunately, although the US government is planning on spending US$2.9 billion on the problem by 1999, it cannot extend checking beyond its own systems and this is where the problem lies, according to Dubai-based computer expert, Mike Kenny. "Some Third World countries use electronic listening devices to alert them of imminent missile attacks but don't have the funds or resources to update. Should these systems be thrown into confusion they are programmed to assume the worst possible scenario and may launch all their missiles," said Kenny.

The problem has been around since 1979 when programmer Bob Bemer, then working for IBM, warned of the possible problems two-digit dates could cause. Bemer had tried to establish an international programming standard for dates and times with ISO but stubbornness, and even laziness on the part of programmers, resulted in two-digits becoming the norm. Former US President, Richard Nixon, also refused to back the initiative, because of his suspicion of computers.

Aside from national security, the biggest disruptions could be caused by malfunctions on big mainframe computers that run institutions like hospitals, banking systems and telecommunications networks. Many of these computers use versions of COBOL, a computer language developed in the early 1960s by computer manufacturers and the US Department of Defense. COBOL was originally set up to use a four-digit date but programmers took a shortcut and used a two-digit format, which is now the standard.

Although armies of programmers worldwide are currently combing key software used by banks and stock exchanges, it is unlikely that they will find every line of code in the next 18 months. Even if they do, the problem doesn't just lie with software. As the price of computer chips has fallen drastically in recent years, coupled with the increasingly sophisticated production capabilities of chip manufacturers, it has become possible to have more software burned directly onto a chip. This means that many electronic devices, from consumer items like VCRs and microwaves to offshore oil rigs use embedded computer chips that store hard-wired information. In this case, the only way to update is to physically replace every single chip. Even if it is at the heart of a nuclear reactor or in a million domestic microwaves. This is particularly worrying for the oil industry, which relies a great deal on real-time programming on embedded microchips. An embedded processor operates as a stand-alone computer within a larger machine. The functioning of real-time software relies on the amount of time consumed between commands. Computers that control automatic machinery must often both detect and introduce time delays of accurately determined lengths to operate machinery correctly. These operational cycles require computers to perform tasks regularly. For example, the command 'perform x task every 2 days'.

One oil operator has estimated that is has more than 40,000 of these embedded microprocessors in its offshore operation. The importance of embedded processors to the oil industry cannot be underestimated. Oil companies should take heed from the fact that the Piper Alpha oil rig disaster was caused by the failure of a single microprocessing chip.

Aside from cost, one of the reasons real-time systems have become so popular in the oil and gas industries is because they produce time-stamped data which is used to generate performance reports and accurately monitor oil and gas supplies. Investigating which systems a company must fix is more important in real-time systems because most only work around one month in advance. If a company does not investigate now, it could find its computer hardware and software crippled at the end of next year.

Unfortunately, updating a single chip may not completely solve a computers' problems. No system failure can be viewed on its own. The knock-on effects of a software breakdown are huge. If there is an electricity power-cut, such diverse consequences could include a loss of a whole city's water pressure, or the complete breakdown of heat sensitive data-processing equipment that is used to store vital information.

But the blame does not lie solely with programmers. A lack of foresight on the part of hardware manufacturers has also contributed to the problem. There are currently several thousand satellites orbiting the earth and almost half of these are more than 20 years old. Even if programmers had anticipated the problem when building these satellites in the 70s, they would not have expected these spacecraft, which in techno-years are the equivalents of dinosaurs, to still be in service.

The reason the problem is so widespread is the importance of a clock as part of the hardware that makes up a computer. A computer logs every transaction and can calculate when to perform certain tasks all based on the accuracy of its clock. Unfortunately, one of the shortcuts many hardware manufacturers take when building their computers is to use inferior quality clocks. One industry insider, who wished to remain anonymous, told Al Shindagah that even the major computer manufacturers refuse to guarantee the accuracy of the timekeeping systems that form part of their products.

So how vulnerable is your home PC to the Y2K bug? IBM claims that all the computers it has produced since 1984 have the ability to handle four-digit dates. But as with mainframe computers, the problem is not simply one of hardware. If a user is running software that is not Y2K compliant, problems could still arise. The best way for home users to check their computer is to visit the websites of the manufacturers of their equipment. IBM has information on their site, as does the American National Software Testing Laboratory. Users without internet access can expect to find a variety of programmes for testing given away free with computer magazine cover disks.

PC users can perform a simple test at home themselves. In Dos type 'date' and 'time' and enter 31-12-99 and 23:59:00. To see if the computer handles the rollover simply restart and check to see if the computer has accepted and retained the correct date.

You can also manually test home equipment by manually resetting the date, wherever possible, beyond 31-12-99. Try it with electronic personal organisers, VCRs and mobile phones. However, users should be aware of the compliance problems with some fixes. One popular technique, called "windowing,'' involves designating a two-digit year value, 30 for example, as a cut-off point. Anything below 30 and the year is assumed to occur after 2000, anything higher and a 19- prefix is assumed. Of course, systems often include a multitude of individual programs that could use different window values, meaning a whole new family of irregularities are waiting to emerge.

The effects of the Y2K computer bug are expected to be incredibly wide ranging. Although there could be hidden benefits. If electronic till systems used in shops can only read up to 1999, any balances will not be carried over to the next day. This could apply to ATM machines at banks. "There is a possibility that money withdrawn from an ATM on 31 December, 1999 will not be deducted from the cardholder's account," said Kenny. This will only work, of course, if the bank's computer hasn't already shut down your account as a result of a Y2K bug in a different software programme to the one used by its ATM.

The scale of the problem has been given credibility by the involvement of America's Central Intelligence Agency. The CIA has admitted there will be numerous failures around the world and is researching the social, political and economic effects the failures could have. Like Kenny, the CIA believes the newly developed nations will be hardest hit. Having invested a lot of money in new technologies, they now do not have the money or the expertise to anticipate and head off impending disasters.

While the US, Britain and Australia have had enough time to handle any possible disruptions many other nations don't have the money or political climate to ensure public safety. In cases where users are unable or cannot afford to upgrade, computer experts believe manufacturers should be held responsible. However, the reason many manufacturers have been slow to deal with the Y2K problem has been their fear of opening themselves up to legal liability for problems caused by hardware and software breakdowns. Manufacturers are unlikely to give written guarantees of their products in case anything goes wrong. However, the importance of dealing with Y2K over apportioning blame is evident in the US, where California has introduced legislation limiting consumers' ability to sue computer companies for millennium glitches.

Although, as one expert stated, this legislation is like making a law preventing anyone from being liable for car crashes, it may be a blessing in disguise. The measure could motivate computer companies to fix problems, while also protecting the economy from a spate of crippling lawsuits.

Although most users believe they have over a year to diagnose and correct any problems, they may be wrong. Many software programmes look one year ahead. So, in fact, 1 January, 1999 may be a good indicator of what to expect.

Another problem to surface is a clause in the system of leap years. The Julian calendar needs an extra day every four years to account for a slight imbalance between the length of a year and the amount of time it takes the earth to revolve around the sun. Because this calendar is approximate, leap year is cancelled every century. That doesn't apply in 2000, which is 400-yearly exception. As if there wasn't enough to deal with!

Right now, experts are looking to 1 January, 1999 for the best possible indication of what is to come on 1 January, 2000. Many programmes that operate one year in advance will demonstrate their ability (or not, as the case may be) to handle the year 2000.

The US and Europe will be keeping their eye on New Zealand, which will be the first country to see in 1 January 1999. With a fifteen hour headstart on the New Year over the US, it could provide vital time to avert the worst crisis. However, the Y2K bug has highlighted a number of other problems caused by programmers' lack of foresight. One such instance is the Dow Jones Industrial Average, which is expected to pass the 10,000 mark in the near future. In a bug similar to Y2K, most systems used by financial centres linked to the Dow Jones are incapable to read figures beyond four digits.As the one Dubai-based computer expert says: "The Y2K bug is not a one-off scenario. We can expect to see related problems well into the next century".