According to early medieval chronicles, the first millennium closed amid widespread belief that the world was about to end. The faithful gathered in Rome-the spot where they assumed that Jesus would return to earth-and prostrated themselves in the basilica of St. Peter's, waiting for midnight to come on Dec. 31, 1999. Many had given away all their possessions, hoping to win a better spot in heaven. As the pope began saying midnight mass, the bells of the cathedral began tolling ominously. Several people died of fear.

After an hour, the bells stopped. The world hadn't ended. Everybody picked themselves up and went home.

You decide.

EDWARD YARDENI: I am an optimist about the Year 2000 issue, and I believe that most major companies will have their computer problems fixed. However, I doubt that we're going to have all the computer systems fixed in time. Maybe 80% to 90% of computer systems around the world will be ready. But surely 10%, maybe even 20%, of computer systems won't be.

The question is: Are these systems vital? I believe the likelihood is growing that the systems that won't be operational in the year 2000 will be important enough to disrupt global economic activity, even to the extent of causing a global recession. I think there's a 40% possibility of a global recession that could be as severe as 1973-74.

I like that analogy because, just as a disruption in the supply of oil caused that recession, a disruption in the flow of information could cause a similarly severe and long recession starting in the year 2000.

I am especially concerned about computer systems run by the government.

DAVID STARR: I agree with half of that. I think government systems are vulnerable. I'm not concerned about business systems for a couple of reasons. One, most of the time when big catastrophes happen in business, it's because of things you don't know. When you know something, particularly this far ahead, most everyone will plan for it. This has probably been the highest-visibility issue in technology in all of my experience.

Now, as to the concern that only 90% of the systems are going to be Year 2000-compliant: Does that imply that 10% will crash and burn? I don't think so. I think the analogy is closer to data security and disaster recovery. Few, if any, businesses have 100% data security. Very few businesses have disaster-recovery plans for everything that could possibly happen, from asteroid collisions to nuclear war.

I think most businesses are asking which systems are going to affect their customers and are fixing all of those. The 10% that don't work should be the Excel spreadsheets and the things that aren't going to have an impact on the business.

YARDENI: Dave, isn't there a risk that what one company views as not very essential to its business and survival is awfully important to somebody else?

STARR: Well, yes, there certainly is. But we have a lot of people who are suppliers to our business, and we have all kinds of inquiries out to them as to whether they're going to have Year 2000 problems, and we're testing their systems. So not only are they testing their own systems and determining which ones are mission-critical, but their customers are doing the same things. So I think you have a doublecheck.

These things are really easy to test. On a 1-10 scale of difficulty, in terms of how hard is it to find this bug and how hard is it to fix it, this is a three. The problem is it's so pervasive that you have to go through this triage process. But my experience so far is that everybody I've been talking to has done just that.

YARDENI: Well, I agree with you. The problem is trivial, yet it is also overwhelming. And I think you need to have 100% of the systems working, or else there are bound to be disruptions. Then it always gets to be a judgment call on whether you think those disruptions will be significant. For example, many bankers tell me they're going to have the problem fixed, but I worry about their loan portfolios. Are all their borrowers going to be Year 2000-compliant?

I also worry about basic service vendors. Electricity suppliers, for instance, are highly automated. Union Pacific is an example of what happens when you have a failure in an important service. Some fossil-fuel electricity plants are reporting that their coal supplies are pretty low right now because they're not getting the coal they need through the Union Pacific line, which has been plagued by computer system troubles.

So I guess my bad-case scenario-I won't say it's the worst-case scenario-is that we have something that's like a combination of Union Pacific, the UPS strike, and the government shutdown of a couple of years ago. Mix it all together, and it's bound to slow economic activity.

STARR: I want to treat the government separately, but most companies-and this is an ugly little secret that the consultants won't tell you-have already had the Year 2000 hit, because people like coal suppliers have contracts that go four or five years out into the future. So this problem is not going to occur with a bang. It's something that's going to phase in system by system over the next two to three years.

YARDENI: Most of my effort really has been focused on the government. And everybody tells me, "Well, that may be a problem." That's exactly my point. We all, one way or another, are very dependent on all sorts of government services.

Look at the phone system. The phone system might be completely functioning, but if it gets overwhelmed with phone calls from people wondering what happened to their disability payment, and what happened to their tax refund, then we could have some serious disruptions.

STARR: I know you're using illustrative examples, but I think the phone system is a good one. They get many more calls on Christmas than they're ever going to get because of this, so they're used to handling those kinds of things. Elevators are another good example. Yes, elevators actually did have a date problem, but what it meant was the automatic call for service was going to be delayed or wasn't going to happen. It didn't mean the elevators were going to stop.

The government, though, is something we ought to talk about.

YARDENI: I've been researching this issue and writing about it since the summer of last year. If you extrapolate the progress made so far, the Labor Department and Energy Department will be ready in 2019 for the year 2000. Defense and Transportation, which includes the FAA, will be ready in 2010.

I am told that the Social Security Administration was well ahead of everybody else because they started in 1989. Odds are that by 1999 they will have cleaned up their 30 million lines of code. Now, apparently, they discovered just last year that they are responsible for another 30 million lines of code at the state disability level. If it took them 10 years to fix 30 million lines, how are they going to fix the next 30 million in two years?

I think KLM and American Airlines have told reporters that they may have to curtail flights in the year 2000 because of concerns about the global air traffic control system.

STARR: What they actually said is they could curtail flights if they thought there was a problem. That's a big if.

You know the dangers of extrapolating. The problem is that government people are assuming that nobody's going to do anything when these systems go down and that we're just going to sit there until 2019. What's going to happen is they're going to get away from the way they're fixing Year 2000 problems now, which is to go through every single line of code. Instead, they'll run some test data through to identify where the problems are, and then they'll fix the bugs. As soon as they're under the gun, they're going to go back to plan B and have it fixed in six months.

YARDENI: I think we're now getting into the gray area of comparative childhoods. Depending on your upbringing and personality, you either have a lot of confidence that the government is going to get it right or is going to screw it up royally.

Are you saying that this is going to be a total non-event?

STARR: No, but I don't think it's going to have any worldwide economic impact. I think there are going to be anecdotal events that will hit the press, so people will say, Oh, Year 2000 happened. But I think everybody will be all over the problem for the next couple of years, and I don't think it's going to really hit business.

The government is another situation. I think it could crash and burn, but I don't think it will take 10 years to fix it.

YARDENI: Well, it if takes one year, I have my recession, I think.

You help run a company that does business on a global basis. Are you as sanguine about what's going on overseas? In Europe, they are distracting themselves with the Euro. Are they going to have the resources? Are they really focusing on the Year 2000 problem at all?

STARR: They're as excited about the Year 2000 issue as they are over the new currency. Because we're international, and our auditors are all over us country by country, we have plans in every single one of our countries around the world.

YARDENI: How do you actually go about finding vendors or programmers? There are so many Year 2000 wannabees out there, companies that claim they can save you overnight.

STARR: My advice is: Don't use them.

The people who know the systems better than anybody else are the people who wrote them. They're your own people.

YARDENI: And your experience is that it is turning out to be less of a problem than you first imagined?

STARR: I didn't think it was going to be a big problem in the first place. But a lot of people I'm talking to are telling me that they're pleasantly surprised it's not as bad as they thought.

It's really not a hard bug to fix. Either you go through every line of code or you put some test data in there and see what blows up. Put something with a date in there that says 2001 and see if the programs stop or if you get funny numbers.

YARDENI: Just on the radio this morning I heard that the FAA did exactly that. They were fixing some sort of software that they had up on the West Coast, and suddenly they lost track of everything in the Pacific Ocean.

STARR: But they checked it, and they were able to fix it.

YARDENI: Yeah, but they were shut down there for a few disturbing hours. [The FAA says that the problem covered part of the Pacific and that it used a backup system to track the planes.]

STARR: Well, that's bad testing techniques. You don't test on your live system. You have a test system.

YARDENI: That goes back to my skepticism about any government agency getting it right. My sense is that it's a big world out there with a lot of computer systems, and, especially now in Asia, they may be so distracted with their crisis that they may not focus enough attention. They might not even have enough financial resources to really deal with the Year 2000 problem.

STARR: I have data centers in Asia, too. My sense is that these problems are compartmentalized. The marketing people are concerned about the economic impacts, but the technologists are worried about keeping the systems up and aren't distracted at all.

YARDENI: I think one of the issues that I have with your analysis is that, while I'm sure you are right on top of the situation, I'm not sure everybody in business is as smart as you are.

STARR: That's my point, though. I don't think you need to be that smart to fix this problem. Anybody who's a competent programmer knows how to fix this.

YARDENI: But management has to perceive it as a problem to devote resources to it.

STARR: I think just the opposite. I think managers have been inundated with terror stories from consultants to the point where they're probably spending more than they should. Almost every week, I get a letter that's passed on to me from some vendor that says the ATM machines are going to start spitting money and the planes are going to fall from the sky.

YARDENI: Well, just to state my position clearly, I have not in any of my writings said that planes are going to fall out of the sky.

STARR: I know you haven't, but that's the kind of stuff I hear.

The other thing that you may be seeing, and I think this is a little more insidious, is it's not necessarily in everybody's best interest to play down the Year 2000 problems. In fact, it's a good way to get rid of a lot of old and creaking systems that you would never have a business justification to get rid of. You can just start waving the Year 2000 flag and saying, "Hey, we need a new general ledger system," even if the payback isn't for 105 years.

The other thing is that it's probably no-win for a CIO to say we're going to save as much money on Year 2000 as we possibly can. If he's right, nobody's going to know it. And if things go wrong, everybody's going to yell at him.

YARDENI: Well, you sound like you're not a fan of the Year 2000 consultants. Wait until you see what the lawyers do with the issue.

STARR: Oh, we've already had letters that were obviously trying to set the stage for legal activities.

YARDENI: Well, we're going to find out who's right. I mean, would you go away to the Caribbean on an airplane in January 2000?

STARR: I absolutely would.

YARDENI: Well, I'm staying home.

STARR: I'll send you a postcard from St. Lucia.

YARDENI: Actually, you might wind up being okay, because I'll be right about the Year 2000 problem, but I'll be freezing here in New York and you'll be sunning in the Caribbean.

The conversation ends with the two agreeing on a bet. Context is to declare a winner in early 2000 and have the loser send the winner a bottle of champagne. Mr. YARDENI can't resist adding: "Assuming you can get through to the loser over the phone system."

Mr. YARDENI can be reached at YARDENI@ix.netcom.com, Mr. STARR at david_STARR@readersdigest.com.