Candle Questions

This section is named in honor of a thoughtful young visitor, who asks some excellent questions about Y2K.  I am going to continue with this page so others can benefit from these conversations.  From this point on, a Candle Question will be a small light that casts a large illumination. I will intersperse other questions and comments. 

Patrick: I am not so much concerned with the effect the Y2K bug will have on the United States, but I am rather concerned with the effect it will have on other countries. I have read several articles dealing with Asia and the economic crisis there. In the articles, they talked about how companies there are having to prioritize their expeditures and the Y2K bug is dead last. My concern is this: If Asia (and other countries for that matter) cannot financially afford to fix the problem and their economy worsens, will it cause a recession, even a depression, here in the United States?

Richard: If the countries are fully aware of the Y2K problem and have relegated Y2K to last place in their lists of priorities, then we can only hope they have chosen wisely and clearly understand Y2K.  As you have already stated, they may not be able to afford to fix the problem.  The questions remain, how much is it going to cost them to repair systems after they break?  And, will it affect us here in the US?   Ed Yardeni certainly is betting on a global recession.  Frances Cairncross at the Economist will probably vote for a recession - with a little extra trouble because of Y2K.  Many people have been waiting for the full effects of the Asian crisis to hit our shores.  It may be that it will be coincident with and combined with Y2K.    Is there anything we can do to help other countries with their Y2K problems?   Do these existing economic and social problems far outweigh Y2K?  I think a recession is likely, but preventable.

Jake:  I recently heard of a talk the local power company had with a group of citizens about the effect of y2k on electrical generation and distribution. We here get very little of our power from hydro electric and rely primarly on coal, natural gas and desiel fuel steamboiler generation.  It has been told that we in this area have a 72 hr surplus of coal, and rely on pipeline and truck delivery (constant) for natural gas and fuel oil.  Do you think this is adequate or should we worry?

Richard:  When I make decisions, I try to compare things in common terms.  Here compare the cost of putting in a supply of fuel again the costs of being without power.  What you need to know is the probability that you will be without power for a certain number of days.  Without knowing where you are, and how severe being without power could be, I cannot advise you.  I spent the last several days reviewing the Gartner Group report to the US Senate in September.  They expect 10% of mission-critical system failures will last longer than three days.  And in the US, 15% of organizations to have mission critical systems failures.  As slow as preparations are going right now, it may be more like 20%.  Take a look at More Questions to see a question Fox wrote about how long might last.  If your power people are on top of it - great!

Ft. Bragg:  In regards to the y2k problem, what do you think is the most likely thing to happen, how should we prepare for the "worst", should we treat this as a hurricane and stock up on supplies, and how will the military be affected?  I am a Private stationed at Ft. Bragg and many soldiers take 2 - 3 weeks off for christmas leave.  What happens if the computers do go down.  How will commanders be able to keep track of all their people, and what if a situation arises and they need to call everyone back to duty quick?  My main question is what should I do to prepare my family and friends.  And how long do you think problems will last?   If all the power systems go down, or the traffic computers go out how, long is it going to last?

Richard:  There are a number of questions here.  Overall, I am optimistic.  Let me break them apart:

• Right now, the most likely thing to happen is that 5-15% of systems will have problems.  The Gartner Group estimates that 15% of organizations in the US will have mission-critical failures, and of those, 10% (~80,000) will have failures that last more than three days.  The Senate report that came out today said that 90% of failures would last no more than three days (~720,000 failures of up to three days).   If you depend on a computer system that was missed, it could take a year or more to repair or replace.  If a city overlooks their traffic control system or emergency communications systems, or emergency vehicles, then loss of life can occur.  Hopefully, any of these systems that are missed will be in the "under 72 hour" category.  I believe most communities understand they should check their traffic controls, community power sources, emergency services, hospitals, and telecommunications.   A step that individuals can take is to ask.
• Many organizations are in a position right now to just miss their Y2K project deadlines.   There could be weeks or months of local chaos as companies switch over to their compliant versions in 2000.  Realize that almost 50% of companies have already seen some Y2K failures as their systems begin encountering dates in the 2000's.  One of my concerns is that many key workers will be taken off preparations to fight fires.  This has already happened to some consulting companies.  Also be aware that many organizations will introduce errors into their systems because of rapid repairs at the last minute, or in the course of very large projects.  These will be discovered in 2000 and beyond.
• In some organizations, it might be prudent to cancel leave over the holidays for critical workers.  In the warfighting business, you know how important preparations and training can be.  Over the next few months, DoD will be conducting some "Table-Top Exercises" to run through some realistic scenarios.  I hope this will not only be a top-level exercise.  Each base should already have groups walking through every critical system. 
• For families, about hurricanes.  The cost to prepare for Y2K will be close to 1 trillion dollars world-wide as the entire stock of computer systems is taken up, checked out, repaired and tested.  The cost of the Kobe Earthquake was US$ 100 bn, the cost of the Los Angeles earthquake was US$ 60 bn.  I think major hurricanes in this country run about $50-80 billion.  Imagine five to ten major earthquakes and hurricanes hitting the country in one year.  I agree with the Senate, several days of food and water - but no hoarding.  Copies of financial statements.

Candle:  If on September 9, 1999 we find that the computers blinking are not Y2K compliant, do you think that there will be such an uproar and panic that the banks would not let people take their money out? I do not mean that they would
close because of Y2K compliancy, I mean that they might fear some sort of depression or stock market drop because people fear that their money is not secure because of the Y2K problem?

Richard:  You are right in that Y2K fears will be fed by significant failures on any of the critical dates during 1999.  If the Gartner group projections are correct, then we will be seeing many failures during 1999 when systems are first faced with dates in the 2000's.  Remember that the failures we will see are only the ones that slipped through the enormous efforts of the Y2K projects already in place.  What we will see is a small fraction of the chaos that would have been if all of these dedicated people had not already been working for so many years.  Right now I understand there might be a slight boom toward the end of the year as people spend more to make their Y2K preparations.  The business activity will be good for the economy.   If too many companies do not complete their Y2K projects, the failures can swamp all our good efforts.  The 9/9/99 error is probably small in comparison with "new dates in the 2000's" errors.

Ryan: "I am a student and is wondering if the Y2K project will effect my going to school?  When will this start taking effect? Is it a rumor about 9/9/99 or is it real? Will this effect School behavior or not?  Why is this so low profiled? People need to know about the disaster heading our way. Thanks, Ryan W."

Richard: I will separate the answers to make them easier to read:

• If you are in a town that hasn't fixed its water, electrical supply, traffic lights, and telephones there might be a slight chance your school would be closed.  Your school building systems themselves - heating, ventilating, security, fire alarms, telephones, computer networks - need to be checked.  I would not get too hopeful, but have your class check to see if your city services are year 2000 compliant, ask your teacher and principle to tell you about the Y2K program for your school and school system.  If they don't have one, there is a very slight chance you might get out of school.  The bad news is you might be in for worse problems.
• Effects were felt in 1969 and 1970 for programs that calculated 30 year mortgages.   Anyone who uses the computer to make calculations that extend into the 2000's might be affected.
• The rumor about 9/9/99 is not a rumor.  This is/was a fairly common programming practice to use all 9's in the date field to indicate a "bad date", "date unknown", "does not expire" or other special condition.  When it was done, it was not considered bad programming practice.  In light of 2000, it is a problem that has to be reprogrammed.
• I would hope it would affect schools in at least two ways.  First, that they check all their programs and equipment and partners to be sure they are all compliant.   Second, that students would be taught "out of the box" a little more.
• The low profile is a combination of things.  It is unwillingness on the part of many people to make overmuch of a simple repair job.  It is reluctance to imagine that such a little error - repeated billions of times - could have such a costly impact.   And, we still do not know of particular "disasters" that will occur.   All of the potential problems, once they are discovered, have been dealt with.   The "disaster" can be a wimper if everyone keeps at the task of finding and repairing. 

Fox: Here are my questions for the day on testing and verification to see if the fixes really do work: Foxheater@aol.com

• Can nukes be tested by isolating a system in a lab, then replace the clock chip so it ticks from 12/31/99; 23:50 then see what happens?  Of course we would hope the nuke would not be armed during the test.  Then can we replace the embedded control chips on this test nuke to see if it functions properly and is safeguarded against self-detonation?
  • Yes.  Each subsystem is reviewed on paper first and tested individually.   Then, with the "nuke" offline, the whole thing is tested.  Not just for 12/31/99, but many other critical dates.  It might not be apparent, but every clock in the system and  connected needs to be set to the same date/time in order for this to be a fair test. 
• Can we take representative systems of major mission-critical to civilization infrastructure areas like power and transportation and do the same tests so we could know now what would happen to our subs, power grid, trains, planes, traffic signals, and other essential to life systems?
  • No and yes.  No there are not representative systems.  The real delay here is that every system is unique.  Certain components and subsystems are "compliant", but collections of compliant systems may not be.  You are right is narrowing in on testing as the essential action in each case.  We need to "prove" that the whole system works, not just the parts and not on paper.
• How difficult would it be to manufacture a chip that has a clock date of 1/1/2000 and use it as a test chip on various systems to see if the so called software fixes actually will work. i.e. remove existing clock chip... replace old chip with test chip... reboot system... see what happens, if the software fixes actually work. 
  • Not too hard, but maybe not necessary.  Some of the chips can have their clocks reset in a test rig.  Some can be set through the software.  The ones I am worried about are the ones where there is no documentation, or where the owners do not feel it is very important.

Candle:  Because there are various time zones in the world, if by chance on January 1, 2000 there were major problems as a direct result of Y2K in places where the new year occurred several hours or even a full day before us, would we be made aware of what happened in those other countries and would it be an accurate example of what would happen to the U.S.?

Richard:  Absolutely!  If you are a multinational bank with offices worldwide, you should be watching for problems as 2000 starts at the international date line in the Pacific, sweeps across Russia, Europe and then the US.  If you are running refinery units in different countries, if you have pumping stations or factories using similar components, you should definitely watch to see what will happen.  A problem I am not sure many are aware of is that most clocks are not set correctly, so that for days before or after 2000 really begins, some machines will make the transition.   Some machines will only fail when they are shut down and restarted for the first time after the transition.  There was some effort in Congress to set up an emergency tracking system to do just what you suggested.  It may not be perfectly accurate, but it will help - as the problem reports coming in now are a help.

Candle:  Exactly how concerned are those most involved with fixing the Y2K problem and
what precautions are they taking? (I read that even Alan Greenspan is extremely worried about what could occur.)

Richard:  The closer you get to the problem, the more concerned you become.  A trillion dollar problem is like five hurricanes and five earthquakes, they greatly disrupt normal life.  Thank goodness much of this cost has been spread out over the last several years and not lumped into late 1999 and early 2000.  But working next to any serious problem, you cannot let fear push aside your will to get the job done.  Y2K managers are, on the whole, confident of their own efforts and rightly so.  Their precautions are to find the non-compliant sections of code and replace them, replace the units that have non-compliant code, do without, or plan to adapt to the problems which come up.  To me a billion dollars is a lot of money, and this is a thousand times larger.  Alan Greenspan was one of those who wrote the code we are fixing now, he knows how hard it would be for him to fix.

Candle:  I've read that Windows 95 and 98 are not Y2K compliant.   How accurate is this statement?

Richard: The term that Microsoft uses is "compliant with minor issues".  As you might know, these operating systems are collections of large numbers of programs and interlinked functions.  In the thousands of things you can do, there are some minor programs and functions that are not Y2K-compliant.  What it points out is that for every program someone needs to look very closely to see that the things you are using in the program will continue to work as you hope they will.   There is a controversy, still, I believe between Microsoft and Intel on who is responsible for the real time clock for Windows NT.  NT apparently reads the RTC when it needs the time.  On a non-compliant hardware system, this exposes NT to problems. Since there are a lot of NT servers in responsible positions, I am wondering if this is a costly problem.   I think I read that there are 150 million Windows users.   If each user has to spend a couple of days going through all their programs to see if they are compliant - expensive! 

Torro: I was wondering whether it was true or not that even though there may be a y2k compliant computer it will malfunction if it comes across another computer that isn't y2k compliant.  Could you verify this for me?

Richard:  The most famous example I can think of was a telephone switch that shut down.  The switch was compliant, but the billing system it was attached to was not.  The billing computer froze and could not complete the transaction with the switch.  Both froze.  Also, your business computer can be perfectly compliant and not be able to work with a non-compliant bank computer.  It is also possible to make a wrong decision in one computer (selection of a premium amount or interest payment) which gets passed through several hands before being discovered.   And, the whole embedded chip arena is filled with stories where small systems fail, causing large systems to fail.

Jill:  I'm hoping you can help me with a question.  Will Y2K in Russia affect the world as I'm seeing on many internet sites?  Some people are saying that Y2K, if not addressed, will cause nuclear bombs to go off.  Is this really true?

Richard:  Like so many possibilities raised, this has an element of truth in it.  There are at least four concerns.  Weakness in our systems will leave us vulnerable, weakness in their systems will make them paranoid and reactive, weaknesses in our partners systems will weaken us, and there is so much to watch we might miss something no matter what.  So, yes, if not addressed, nuclear bombs could go off - which seems to be the point of the actions by Hamre and others.  I feel most of the obvious problems are being addressed as people uncover them and press to get them corrected.  As long as there is a non-zero probability of an accident, and we have time, let's put some effort to finding out.   The consequenses of not checking are grave.

Please take a look at my notes on some of the material I reviewed to answer your question.  I am starting a section on Defense.

Candle: I heard that because of Y2K , Washington may have to shut down our national security (like the places where the missles are stored) because if they let it remain powered up, it could cause an accidental nuclear war.  How far fetched is this idea and exactly how possible is it that this could occur?

Richard: The plans now are not to shut down the systems, but to have a joint US-Russia early warning team to track events in both countries.  Working together, they will provide a human channel working to prevent accidental escalation.   That center has not been established.

Candle:  I wanted to let you know that there is an excellent article in the January 1999 issue of Vanity Fair about the Y2K problem. It is about 8 pages long, but it is quite easy to understand. That may be a good item to put on your website in case someone is interested and it's very comprehendable.

Richard:  The article, "12.31.99 - The Y2K Nightmare" by Robert Sam Anson (page 80) is well done.  From the title the reader will hopefully gather that this is partially intended to point out some of the more sobering aspects of the problem - questions of life and death, personal and national survival are intended to get our attention.  I found the background on some of the people in the early days of Y2K to be fascinating;  but, unfortunately, I do not have the time or resources to confirm these details.  I can only say that it does not conflict with what I know.  The focus on the people is refreshing, to give some background on why this happened the way it did.  At this point VF does not post its articles online, so I cannot refer visitors to the article.

Fox: I read your response to Candle, and I've been witnessing the evolution of the public and television consciousness to the Y2K delimma - maybe we can make it after all!  I really hope so!  If our government will concentrate on  those areas that support the mission-critical infrastructure, everything else will take care of itself, even if it means going back to receipt pads, etc.  As long as the banking system, power, and nukes are under control, everything else can be overcome!

Richard: Yes, there is reason for hope.  But, each organization, each individual needs to be aware of the computers that affect them.  Y2K is not a difficult problem, but it requires diligence so no individual is harmed.

Candle: I keep hearing: "Oh, America is too smart. We have brilliant people that can figure this out in no time. Hey, its only 2 stupid digits. What could be so hard?"

Richard: I know of thouands of people working on the problem full time.  The cost in the US is going to be about $600 billion dollars - just for the fixing ahead of time.  Among the people I know who have to deal with it, it is a tedious problem. Yes, it is not rocket science.  There are just a lot of computers and a lot of lines of code accumulated that need to be fixed.  We have had a whole generation of people working very hard to write all these billions of lines of code, install these millions of PC's, design and install these millions of embedded systems.   8-10% of the lines of code need to be changed ($1 to $5 per line).  Half the PC's are non-compliant and much of the software (3-4 major packages per machine) is non-compliant.  5% of the embedded system have to be evaluated and 0.1% of them need to be replaced because they will fail.  Does it all need to be fixed, no.  But every system will have to be considered - because we forget what we have, because we forget how depend we are on things, because these systems are so interdependent..

Candle: I also hear:  The "simple stupid 2 digit problem" will be solved before the year 2000 arrives.

Richard:  This problem has already been solved.  Or at least the bulk of it.  The answer is Find it - Evaluate it. - Fix it.- Prove that you have fixed it. - Plan what to do if you make a mistake or miss something.  But, there are so many systems.

Candle:  I was wondering exactly why the public hasn't been made more aware of this problem? Do you think that Washington wants to conceal or downplay the problem for as long as they can in order to quell a would be panic?

Richard:  Yes.  About 5% of the population is in panic mode, 30% are in denial, and the rest are just going about their lives.  Every manager of a Y2K project that I know of is taking a serious and diligent approach.  But many organizations have not even started.  Yet more and more communities are considering disaster preparedness drills.  This is why I work so hard to try and find out what is really going on - to keep people from spending too much time and effort preparing and practicing for things that are not likely.  At the same time encouraging engineers and designers and people across the board to dig harder and faster now, so we can find any serious problems.  Every step is hard - Find it.  Evaluate it. Fix it.  Prove it.  Prepare in case you missed something.  I have read tens of thousands of pages of material on this.  Let me tell you it is serious enough to address - steadfastly and without panic.

Candle:  ...while probably about 99% of our nation is skeptical that anything will happen.

Richard:  Things are already happening, but mostly where people are prepared and watching.  About 46% of corporations have already had Year 2000 failures.  The number and severity will increase over this year.   But at the same time only a small percentage of systems are going to fail - it is just that there are so many of them.  (I highly recommend you read the Gartner Group report to the the US Senate.  Look in International Organizations at the top).

Candle:  My other two questions for you has to do with the triple nines and the quadruple nines. Exactly what is the significance of these numbers? I have heard that on September 9, 1999 it will display some sort of warning or indication that the problem is not fixed and that the same will happen April 9, 1999.

Richard:  There have already been failures reported where companies have missed where a programmer used 99 as "unknown year" or "invalid entry" or "expired".  These YR=99 and 9/99 and 9/9/99 problems were uncovered and well known for years as programmers have been fixing date problems in systems all over the world.  But, there are a lot of programs and programmers out there.  Remember that onion I mentioned.  Every organization that has done a serious survey finds more and more and more things they did not know they even had.  And often it is the obvious, after the fact, little, stable, quiet, always-runs, no-problems, yet mission-critical systems that you might miss.

Candle:  Do you think that if there was some indication of a problem at one of these dates that they would close the banks because people would panic and run to take their money out? Personally, that is what I think, but I was just wondering what your feelings are on it.

Richard:  On the 99th day of 1999 there is going to be a problem somewhere, because somebody missed a piece of old code.  But every bank does not use the same software, every manufacturing plant does not rely on the same equipment with embedded software, every telephone company is different, no two community sewage treatment plants are identical.  Our diversity and variations are a strength when it comes to preventing failures.  It is not likely that a common failure will propagate across all systems.  But this same diversity of systems makes for a horrendous number of systems to go through.  So the finding and evaluating and proving become harder.   What I am trying to do is to look for what we might be missing.

Candle: (Asking what is going on, and should a family prepare for Y2K)

Richard Collins:  Hello. Thanks for writing. Even though I spend all my time trying to understand this Y2K problem, there are so many unknowns it is still hard to predict what will happen. We recently went through a two day power outage because of an ice storm. We about froze! So I do not think that having a stock of basic necessities is wrong-headed at all.

• A friend in Mexico City says she is moving out because of Y2K (there are 20 million people in Mexico City). Many countries are not moving quickly enough. At this point many companies have not started any kind of Y2K program. I think they are going to be unhappy when they find out how long it takes to do it. Most people have no idea how many computers they use.
• At http://www.eei.org/EEI/press/y2k/members.shtml   you will find many electric companies with Year 2000 programs. They are treating it seriously.
• I have been meeting with people in the Transit industry and Defense. They are moving, but not as quickly as I think they should.
• All in all, things are improving. But, here is a joke for you. "Y2K is like an onion, the more layers you peel back, the more you want to cry." It is a surprisingly difficult problem.
• You and your parents can help by at least asking your local electric, telephone, water and emergency organizations about their Y2K programs. This is something that even you can do with your class as a school project.
• You asked about some links to sites to help in preparations.  Here are some pages I put together for you.  I hope to have time to do more:
  • Community Preparedness (draft)
  • Community Preparedness Groups