To: Boob who wrote (2571 ) 5/25/1998 12:11:00 AM From: Ga Bard Read Replies (2) | Respond to of 9440
ALYA ... This is an e-mail I think everyone considering or owning ALYA should read. Y2k companies are hype not because the problem is easy to solve, but because it is insoluble. You are right that the only way to fix it in many cases is to junk the system and start all over again. Silver bullets and the like are nonsense. The only way to go after Millennium bugs is to go through gajillions of lines of code. There aren't enough people remotely trained enough to do the job. It's easy to miss something, and then your ostenisbly de-bugged product isn't debugged at all. And then you have the problem of chip logic. Many of these are produced by Taiwanese bucket shops from pirated specs, or are otherwise reverse- engineered. Makers of products from toasters to elevators then buy these chips from catalogs. Many of these chips in their present incarnation have no proper documentation at all -- and a lot of them have timers built in, which may or may not be accessed in the actual product in which they are embedded. But take heart: tax compliance will be voluntary in 2000. The IRS is going down. If you don't believe me, just watch Congress. Why do you think they're doing the big PR campaign about curbing IRS abuses now? Hoping Joe Mainstreet will feel more cuddly about the taxman, and make sure he sends in all he owes in the new Millennium. A snowball's chance as far as I'm concerned. But it shows how desperate they are. The I.R.S. is in incredibly bad shape because its hardware is antiquated (many computers still on-line that date back to the early 70's. They still have Univacs, if you can believe that). The software has been patched a billion times by different hands to reflect changes in the tax code. It doesn't quite work now, and to the degree that it does, no-one understands why. The 6 gurus they have supposedly hired cannot do squat to fix this problem. 600 crack programmers could make a dent. But you cannot find them. I am in a Masters in Computer Science program at NJIT, the biggest such program in the NYC area. (Yes, I'm 46 years old and have computer-related patents. Anything wrong with learning more?) I do not know one graduate student who is capable of being of help. Guys who can do this stuff get $250-500k per year. The government has no budget for this, and couldn't lure them away from private industry if they tried. The magnitude of the job and the deadline demands 6000 programmers, but there is no known way to coordinate a team that large. Don't expect the government to be the one to figure out how. Private industry will figure out how to deal with Y2k reasonably well. The Stock Exchange will be fine -- the best talent in the world is working on it. Big Banks should be ok. Smaller ones might not. But tracking this stuff through all the nooks and crannies is near impossible. Some supposedly- debugged systems will have problems anyway. And all of these systems interact with one another through networking in ways that simply cannot be predicted. The real problem is not that you will get a bill for 100 years interest on your mortgage. Haha, very funny, good one-paragraph filler for the Readers Digest. The real problem is that computers freeze when they get inconsistent data. So you will get no bill at all. It is guaranteed that some of the 50 states and who knows how many counties and municipalities that distribute welfare will not be able to produce the checks. The National Guard will have to be sent in with sacks of cash to avert riots. Since we have fiat money, this is not really a problem -- at least short term. The government can print as many greenbacks as it needs. Some effort will be made to keep track of what's happening on a million PC's, which should be compliant. But they will not be all networked -- there isn't time to build a parallel system capable of tracking the government's business, even if you started it ab initio to escape the problem of lousy data from the past. And let me tell you something. Nearly every database in the world has dates in the mm-dd-yy format in permanent storage. The incalculable problem is to figure out and trace the unwritten assumption about the century which is present in nearly every program that crunches that data -- and which may be different in different modules or layers of the same "program". A further problem is the Third World. There will be 50 Indonesias as their central banks go down. The low end Third World nations have donated 70's and 80's computers that were cast-offs then. Not one national knows how to take the cover off. Nobody is about to pull people off the Stock Exchange and send them to Chad to make sure everything will be all right. The big boys are planning to create as much paper and digital money as is needed to plug all the holes in the dikes, even if they can't account for it properly. It will require trillions. As that float makes its way through the system, the Volcker-Greenspan era of low inflation will come to an end. Oh, yes, Alan's term is up in June 2000. He will be replaced, not with another superannuated Ayn Randian gold-bug hard money nut, but with someone who will do exactly what they want -- and frankly, what needs to be done. The Y2k threat is not a stock market crash, though the market will be uneasy, and scary declines are more than likely. (Our sector will survive better than anything else). The Y2k threat is not recession or depression, as the scaremongers would have you believe. The Y2k threat is inflation, which will take hold in 6-18 months after Y2k as a result of what they did to deal with Y2k. There's a lot more that I could say, as I've made this whole area one of special study. But this will do for the broad outlines. This scenario is as solid as your DD, which has richly benefiited me in SURG, MIDL, NUTK and others. Keep up the good work! Now that was a solid e-mail of an intelligent researcher. GB