The Y2K problem is helping mainframe sales, esp. for IBM. Many companies are deciding that now is the time to update their mainframes or scrap them in favor of smaller computers. Their existing investment in mainframes often wins out, especially since replacement machines are much improved and no longer require expensive computer rooms. The market for mainframe databases and analysis software is expected to jump from $5.5B this year to $22B in '02, and the overall [mainframe-related software and e-commerce?] market from $23B to $113B. Oracle, Sybase, and Informix have concentrated on the Unix world, but IBM has retained its mainframe database focus and DB2 customers. [Michele Hostetler, IBD, 19Aug98. , comp.databases.ibm-db2, 03Sep98.] (But does it really help the Y2K problem to be switching computer systems while trying to debug or rewrite legacy code? Sounds like a recipe for disaster. Then again, keeping the old hardware and software doesn't sound much better.)

For more about Y2K -- on the scary side -- see the "power_grid" and "non_compliant_chips" sections of Gary North's web site at . Date functions are widespread, such as in event logging and process timing. [, comp.software.year-2000, 03Sep98.]

One of the optimistic articles I've seen -- though less so than most articles by journalists -- said that we will survive by shutting down everything not known to be compliant, then bring up and debugging systems one by one as we have the time and urgent need. We can run a skeleton operation for as long as necessary, like the Enterprise running life support until the engineers can bring other systems back online. Critical systems are the ones most likely to be compliant by 01Jan00, or very soon after. As for the rest, we're likely to find that much of our billions of lines of code aren't really necessary. (The US government shut down for a month last year; was that critical?) And if some companies can't keep vital services running, business will shift to alternatives. If your inventory program doesn't work, buy another that does -- or keep track of inventory with pencil and paper.

On the other hand, the 300 working days left are hardly enough to fix and test even the more important software applications, let alone replacing process control boards and other embedded systems. Most of the comforting spin being given to the press comes from executives who are far removed from any actual code or hardware. The remainder comes from programmers who have completed their Y2K work. For every such programmer, there are N others who won't talk because they'd lose their jobs, and there are uncounted conversion tasks that have been started or even considered. Besides, software projects always take several times as long as predicted. Executives all say they'll be ready to begin Y2K testing by 01Jan99, but that's just words -- they can't be honest because of the lawyers, or because they'll be fired. In Australia and Canada, it's rumored, the military are preparing for civil disruption. You don't see that in the US because it would be politically risky. We may find that delaying the bad news only heightens feelings of distrust and panic when systems start falling apart.

Facts are hard to come by, and hard to trust when the person citing them is also selling doomsday books and lectures. Here's one data point: The California Dept. of Motor Vehicles brings in $3B in revenues each year. Is that a critical function? This agency that had to scrap a massive computer upgrade after spending enormous sums over many years. Anyway, Tech Week for 10Aug98 says that the CA Dept. of Information Technology (DOIT) estimates it will cost $240M to fix all the state agencies' Y2K problems, with only $55M actually budgeted this year. The CA Legislative Analyst's Office warns that DOIT's estimate excludes embedded chips, underestimates costs, has overly optimistic time lines, and lacks contingency planning and oversight. [, comp.software.year-2000, 11Aug98.] If state and national cash flows are disrupted, what effect does it have on budgets for Y2K work? Ditto for corporate problems. If services are disrupted, what effect does that have on society? If paychecks stop, what happens to people who are living hand to mouth? And to the stores that serve them, and to the people working in the stores, ad infinitum?

Even if we survive the [brief?] loss of major governmental and corporate functions, what will happen to the stock market? And what about bank runs? If people want their money in the mattress, they don't want it in stocks. That's bound to drive the market down, in addition to any effect from collapsing economies worldwide. (Many are on the brink now because easy credit and capital from the past several years are being withdrawn. That affects all importers and exporters, those who buy or use imported parts, and those whose jobs depend on export.) I've heard that 1929 can't happen again because we have the SEC riding herd, but I don't believe it. (Not that I know anything about economics or banking -- although my paternal grandfather made his living closing banks during the Great Depression.) Even the SEC says that most companies have issued only boilerplate Y2K disclosures, just adding a bit of legalese to their financial reports. [Los Angeles Times, 03Aug98. EduP.]

That same Tech Week article quoted Senate testimony that 38% of IT professionals say they may withdraw personal assets from banks and investment companies just prior to the millennium." (.) The US financial industry itself is in relatively good shape, but it's vulnerable to problems in all other industries and countries. ("49 major Japanese banks planned to spend $249M as a group on Y2K compliance. This amount is only a fraction of Citicorp's planned $600M program.") Even US financial industry compliance has never been tested, except for a limited stock market test of some testing tools to be applied next spring. (.)

I don't think we're going to see Road Warrior survivalism, or even village culture and banditry, but I suspect we're in for a hell of a recession. What I haven't figured out is whether it will be accompanied by inflation or deflation. Bank reserves total only $53B, which won't go very far if 100M people start pumping the ATM machines for two month's living expenses. The government is planning to shovel out an extra $200B, mostly in large bills. That would be inflationary, but it's deflationary if there's not enough money to go around (due to runs and hoarding) or if people stop buying goods and services. During the Great Depression, those with money could hire those without for $1/day. That's deflation. Mortgages were foreclosed; people lost their houses. On the other hand, government could shovel out money until it's completely worthless. Savings would evaporate, but debts could be paid off with worthless paper. Barter would be the new currency, with wealth consisting of food, bullets, seed, plows, salt, coffee, etc. Ever hear of that happening to a country?

The government is doing worse on Y2K remediation than any other US sector, but it's government's job to pull us through no matter what. That's the reason for setting up emergency powers, as in Clinton's Executive Order 12919 and others (). The EO gives FEMA and other government agencies the right/duty to confiscate commercial or private food, water, generators, tools, transportation, etc., during times of emergency. Department heads are authorized "to employ persons of outstanding experience and ability without compensation" (Sec. 602), in addition to other means of employment, and the government will "develop policies regulating the induction and deferment of personnel for the armed services [for] the labor requirements of defense and essential civilian activities" (Sec. 701). The government can also "require acceptance and priority performance of contracts or orders ... and allocate materials, services, and facilities as deemed necessary or appropriate to promote the national defense." [Dean T. Miller and Jim Abel , comp.software.year-2000, 13Jun98. Bill Park.]

Reassuring, I guess, although libertarians find it alarming. BTW, the US Navy recently took two battle cruisers out of commission while they integrated two new weapons control systems with the ships' legacy software. One of the upgrades took the Aegis Baseline 6 air defense system -- written in Ada, C++, and several other languages -- from 3M lines of code to 8M. The new systems worked individually, but wouldn't interoperate. (. [, comp.software.year-2000, 17Jul98. Bill Park.]) Not unusual, considering the trouble that other agencies have had in upgrading their computer systems or the delays of commercial companies in bringing out new operating systems and applications. These aren't jobs for amateurs, and you can't bring programmers up to speed in a couple of days -- esp. for Y2K, where problems may be distributed all through the code. That's why the Y2K programmers and king geeks in the UK are getting #400-#900/day, or about $165K-$375K/year in US terms. [Michael McCormack, comp.software.year-2000, 18Jul98. Bill Park.]

In the US, lawyers are the ones expecting windfall wealth. A CA judge recently ruled that companies can't be sued in advance for damages (thus raising funds to mitigate the problem). [NYT, 07Sep98. EduP.] We can expect to see many vendors put out of business by legal action if not by economic chaos.

There's plenty more on the Web. See MITRE's for one of the best detailed overviews of Y2K technical challenges, or for a discussion by Y2K experts. NIST also offers Y2K tools and information, at .

For the darkest views, read "Alas Babylon" or other post- nuclear fiction, or maybe Salisbury's "The 900 Days" about the siege of Leningrad. I wonder just how long it would take for New York City or San Francisco to die out after food deliveries stop, or how long it would take to get the cities functioning again. More importantly, how long would it take Palo Alto? But maybe we don't really need cities or newsletter publishers.

-- Ken