Wrestling with Y2K

Y2K, The Millenium Bug-- you've probably heard of these, and if you haven't, you undoubtedly will over the next year or so. Some people think it's going to be the biggest problem to face society since the Great Depression and the World Wars. Others expect it to be a minor speedbump that you might miss if you aren't watching the news that weekend. In any case, some amount of preparation, both mental and physical, is a good idea. The more people that are aware early on the fewer problems there'll be.

This essay is longer than my usual ramble, so I've broken it into sections:

• THE BASICS -- 1/1/00
  The core of the problem, as well as some of its ugly cousins.
• BUT WHAT DOES THIS MEAN TO US?
  How the problem could reach you.
• OPPOSING VIEWPOINTS: THE GOOD, THE BAD, THE UGLY
  What others are expecting, as well as some useful links.
• WHAT TO DO
  How to react to an issue so big.

Some programmers have always been aware that there would be a problem when the year stopped climbing from "97", "98", "99" and dropped to "00". At the time however, few managers believed that the systems they were responsible for creating would be around that long without being replaced. (And if you think that's shortsighted, on some projects it made sense to use one digit years, and they had a "Year-1980 Problem" instead.) Also, using a two digit date comes naturally to people. When we talk to each other, we're smart enough to know that if I say I was born in '74, I probably don't mean 1874. Computers lack the common sense to make the same logical leap, but programmers might not have always had that logical gap in mind.

So what happens when a computer tries to work over the transition from 12/31/99 to 1/1/00? It depends. Let's say you ask one of these systems how old a man born in '54 will be in '00. It might happily do the math: 2000-1954=46. It might just as happily think 00-54=-54. In some odd cases, it might try 19100-1954 = 17146! In many cases, it will give the answer "SYSTEM CRASH", and that might be the biggest problem of them all.

The other issue is that humanity will always be dealing with these kind of date failures: older Unix systems will fail in 2038, some chips use day counters (instead of human readable dates) that could fail at arbitrary times, and who on earth is preparing for the year 10,000 problem?

To understand the potential size of the problem, you have to recognize a few things:

• Computers are everywhere in our society. From agriculture to manufacturing plants to warehouses, from businesses to government, from the transportation system to the electrical power grid to the telecommunications setup, computers keep track of records, print up schedules for people to follow, balance situations that are too complex or too boring for humans to handle day in and day out. We could not support the number of people that we have in urban areas, far away from the sources of food and raw material, without computers. And before you say "but we had New York City before we had computers," keep in mind that many techniques for supporting populations without computers are likely to have been forgotten over the last few decades, replaced by computers systems that were becoming cheaper and more efficient than humans doing the same job.
• There's a variation of Murphy's Law by Douglas Hofstadter (a computer scientist) that goes

  "Hofstadter's Law: Projects always takes longer than you expect, even when you take Hofstadter's Law into account."

  --this law is especially true when trying to fix or write a computer program, especially ones as complex and mysterious as the ones used by large companies. Overruns of months and months and millions of dollars aren't uncommon. There are even programs that have been quietly working for decades, causing everyone to forget how their internals work, or for which the records for the code have been lost.

  To people who haven't programmed, it might seem to be a matter of finding the one solution and using that everywhere. There is no one "magic bullet" that will work in this way. Many of the computer systems are unique, with their own histories, their own tasks to do, and their own ways of doing it. Even worse, there are literally millions of programs that need to be at least looked at, if not in need of complete overhaul. Jim Lord (I reference his website below) compares it to polishing marbles: each marble that needs to be polished might not be difficult by itself, but we've got a Grand Canyon full of these marbles, all of which need attention. I think this isn't a perfect metaphor, since the marbles are all different and the polishing is far more important for some marbles than for others, but it gives you an idea of the scope of the problem.

Some of the more troubling parts of the Y2K bug are the problems with "Embedded Systems". Unlike the old mainframes that people might interact with on a regular basis, these are computer chips sitting in remote (and not so remote) corners of millions of pieces of equipment: lighting systems, security systems, elevators, railroad crossings, medical equipment, possibly even pacemakers. The good news is that the lion's share of these systems have no clue what the date is and shouldn't have a problem. The bad news is no one has a clear idea of whether any given system will have a problem without doing a series of time consuming and expensive tests. The worse news is that the tests that have been done have shown mixed results at best (A story is circulating that a GM plant rolled its clocks forward and every robot arm (and its entire manufacturing system) just stopped, frozen.) The even worse news is that some of these chips are going to be very, very difficult (in some cases, impossible) to find or make replacements for, particularly before the looming deadline. In some cases, you can fool the system by rolling back the date, which may work but creates headaches of its own.

Even if you don't work for a company that uses these processors, you depend on electricity being steadily provided. The electric grid is a vast network that can barely support loads during times of widespread air conditioning use. If they had problems getting fuel supplies via rail or truck, or in trying to make up for nuclear stations that were forced to go offline for safety reasons, the entire system would be hard pressed. Water and sewage systems, even if they had no Y2K issues of their own, would not work without electricity. Similarly, the transportation system (often running near maximum capacity, and with many Y2K issues of its own) is depended on by almost every other sector of the economy. If the warehouse can't get the materials from its supplier, the local business can't get its product from the warehouse. No one knows for certain how robust or fragile the chain is.

On the one hand, Y2K (hopefully) will not have many of the large problems wars and disasters provide: people will not be running from sniper fire, they will not be trying to dry out from flood conditions. On the other hand, in all of these situations there has always been an unaffected "somewhere else" to send in relief and help with the rebuilding. With Y2K there will not be a completely unaffected "somewhere else."

Some people believe that Y2K will be a fundamental stumbling block for Civilization as a whole. Some of these people have ties in to Radical-Right Wing politics and come to the debate with an agenda-- they wouldn't mind seeing the end of the current system of Federal Government and expect to see cities aflame. Others in this group are out to make a profit with Survivalist goods, consultant services, or books of their own for sale. Still others of these people have no such agenda, are very intelligent, and are very concerned. The pinnacle of these sites is Gary North's Y2K Links and Forums. Gary North has a huge collection of articles with all links prefaced by commentary by him. It's a very scary read, though I believe his mindset is such that he will usually put a negative spin on things in his commentary.

Yahoo's Full Coverage has a reasonable number of headlines updated on a daily basis.

• Become aware of the problem, and help others to become aware. There's a lot of skepticism out there, and much of the skepticism isn't warranted. Convincing people that many things will go wrong is a large part of the battle.
• The other part of the battle is to prevent people from panicking. Even if not a single computer ended up having a problem at the rollover, there would be economic problems if people decided to convert their savings into cash and shortages as people wanted to stock up on food and survival gear. The more people panic at the cost of cooperating, the worse it's going to be. Running for the hills may sometimes sound like the safe bet, but unless you've had years of preparation in fending for yourself you'd be in much more danger than if you'd stayed and prepared at home. (Especially considering that the efforts to fix problems with the infrastructure will focus on the cities first.)
• Prepare. Many people of moderate opinion think that it would be wise to have a supply of food and drinkable water on hand, just in case. You may wish to stockpile prescription medicines if possible, and maybe have a little extra cash on hand (though probably not your entire life savings.) Make an effort to know your neighbors so you can help each other out, as well as getting the word to that many more people.
• Learn to keep a level head as you keep yourself informed. When some writers speak of "disaster" they might mean something that will have a harmful effect on people's standards of living, not on their survival. If enough things broke without warning, survivability would become an issue-- but we've had warning, priorities have been set, and by the time the date does roll around only the very isolated will be uninformed of the issue. Life may well be very different and somewhat less comfortable for a number of years, but people are tough and smart, and it's my current opinion that Society is bigger than this problem that it faces.

(For the record, I prefer to write dates in YYYY/MM/DD format, such as 1998-8-30, since it avoids the Europe/ USA order difference, can be used to sort dates in correct order, and just looks so darn cool.)