KoSxxT's "judegment". Here's the Webster/InfoNOmagic debate KoSXXt gave to "infoNO"greenspun.com : LUSENET : TB2K spinoff uncensored : One Thread |
-------------------------------------------------------------------------------- Uhh, nice Hoffy, very nice....For a serious debate about Infomagic's position, the one that he had with Bruce Webster on adaptive systems in one of Cory's newsletters is great reading. (Webster got trounced, IMHO.)
-- King of Spain (madrid@aol.com), August 24, 1999.
xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
buttttt. first a word from the sponsor
This article is published as part of Cory Hamasaki's DC Y2K Weather Report and may be reproduced under the same terms and conditions. All other rights reserved to the author. Preparedness -- Press Release -- The Y2K Survival Supply Co. is providing affordable freeze dried and dehydrated foods, water storage tanks, water filtration units, solar electric systems, generators, and many other useful emergency and survival products. Several of our one-year food storage units can be shipped to you in as little as three to four weeks. Solar panels, inverters, chargers and other devices needed for your solar electric system are in stock and can be shipped immediately. If you mention the WRP's you will receive an additional 5% discount off our already low, low prices!! This offer ends February 15, 1999. Time is running on and the only words that make sense are DON'T DELAY - PREPARE TODAY!! Please visit us at - www.y2ksupplyco.com - or call toll free 1-888-925-9258 Intro -- Bruce Webster and Infomagic on Charlotte's Web -- An early draft of Bruce Webster's thoughts on Infomagic's "Set Recovery On - Part II, Charlotte's Web" (see WRP 103) escaped to c.s.y2k and the Internet a few weeks ago. Bruce has finished the article and has allowed the DC Y2K Weather Report to run the final version. Infomagic is responding directly to Bruce's critique. Both articles were issued in text form to Internet Newsgroup Comp.Software.Year-2000 in the week of December 16, 1998 as a prepublication preview to WRP 106. Both Bruce Webster and Infomagic have significant Enterprise Systems expertise as well as recent Y2K remediation experience. Please take the time to read their essays, pass the information to others, and discuss and debate their ideas. Commentary -- Bruce Webster -- Comments on Infomagic II by Bruce Webster Many of the Year 2000 postings paint a picture of significant or even irreversable collapse should there be significant outages in the "iron triangle" of power, telecom, and banking. Their arguments are that such failures will inevitably lead to a domino effect, pulling down system after system. At the very least, they argue for mass bankruptcies, especially among large businesses, complete shutdown of air traffice for extended periods, and so on. They cite the inevitable unfixed Y2K bugs, as well as the iatrogenic defects created when attempting to fix Y2K bugs. [The term "iatrogenic" is a medical one and means "doctor-caused". An "iatrogenic" program defect is one introduced by a programmer trying to fix an existing defect.] Such analyses go wrong in several areas. First, they often show a lack of understanding of key business issues. For example, the most likely reason for diminished or interrupted commercial airline is not FAA problems or problems in the aircraft themselves or even in the airports. It is that insurance companies will withhold liability insurance until they are certain there are no bugs in the system, so to speak. As someone else has noted, could you imagine what the repercussions of the Swissair tragedy off Nova Scotia would be if it had happened on January 1, 2000? Some analyses reflect a lack of understanding of the realities of information technology. For example, one analysis asserts that even if we should fix all Y2K problems (which, like others, I consider an impossibility), the residual failures alone would cause a global economic disaster. This ignores the vast number of software defects currently present in the global IT infrastructure, which we manage to work around somehow anyway. It also asserts that such defects are "systems failures" as opposed to annoyances, miscalculations, and so on. Let me give you a counter-example. I know of a major commercial enterprise that estimates that it loses over $100M a year in overpayments to other firms due to software errors and inadequacies. Yet this firm is profitable and sound. Many analyses talks about businesses failing without projecting a clear understanding of how businesses fail. In the vast majority of cases, it's due to a lack of cash (or credit). Note that businesses, especially large ones, can suffer major losses for months or even years (cf. Apple) and stay in business. These analyses likewise fail to cite situations where a software failure has ever been fatal for a business, especially a large one. Again, most software, especially that within enterprises, is defective; failures occur daily and are dealt with. There are a few real-world examples of software failures crippling organizations, as can be seen in the books Software Runaways (Glass) and Software Failure: Management Failure (Flowers), but what those books more emphatically point out is the ability of enterprises to absorb tremendous software failures (and financial impacts) without going out of business. Such analyses need to better establish how Y2K failures could lead to such business failures. Most Y2K failures range from minor annoyances to major inconveniences. Some could seriously hurt a given business. But the analyses do not give credible scenarios where Y2K defects (much less residual or iatrogenic ones) would lead to a Fortune 500 corporation going into bankruptcy. The supply chain argument is the most credible one made. The GM shutdown this year due to a 57-day strike at two suppliers led to 200,000 workers being laid off, while GM lost nearly $3B. But note: GM didn't go out of business. Likewise, assertions that large businesses are more likely than small businesses to go bankrupt ignores the realities of business: small businesses are far more likely to close than large businesses because they are more vulnerable to shutdowns in cash flow, supplies, and services. Finally, the arguments about inevitable collapses, devolutionary spirals, and carrying capacity are nothing new. In various forms, they were at the core of books such as The Limits to Growth (Forrester), The Population Bomb (Erlich), and The Coming Dark Age (de Vaca), all of which turned out to be dead wrong in their predictions. Why? Because they presumed that the complex systems--social, political, economic, informational, technological, and so on--that make up human civilization would fail to adapt or adjust. They also ignore the absolutely wretched state most of the world was in just 50 years ago after World War II--with most of Europe, Russia, and Asia laid waste, hundreds of millions dead, crippled, sick, or undernourished, occupying armies spread throughout the world, and the US economy the only robust one on the planet--and the subsequent economic expansion which has taken place in those 50 years with very few bumps (1973-74, 1980-82, 1991) in spite of--and in some cases, because of--that devastation. From what I can tell, those making such arguments are largely unaware of complexity theory and complex adaptive systems (cas). I don't profess to be an expert on complexity theory, either, but I've studied it enough to see how enmeshed its principles are in the world around us, including--especially--the human side of that world. Very briefly put, complex, intelligent, adaptive behavior emerges out of a cas not because of some form of central control or mass goals, but because of simple rules followed by the agents who make up the cas. As noted above, we are enmeshed in a vast, interlocking network of complex adapative systems which we do not fully comprehend and certainly do not control, though as agents we influence and shape them. It is my personal opinion that the 'miracle' economy that we have enjoyed for the past several years is a result of emerging complex systems. If you went back six or seven years ago and asked a group of economists whether it was possible to have a national economy with low inflation, low unemployment, low interest rates, and steady growth, you would have been laughed out of the room; it violated virtually every school of economic thought. Yet that's what we've had for some time now. Lest you think that complex systems are your unconditional friends, recognize that they have consequences and side effects that we may not like. As consumers, we are (generally) thrilled with the stunning drop in the price/performance ratio for personal computers; a four-month delay in getting a new system has saved me 25% or about $1000. But I sure wouldn't want to be Compaq or Dell or Gateway or CompUSA; the competition is fierce and the margins paper thin. Likewise, the growing rash of mergers and acquisitions, with commensurate layoffs, are likely symptoms of how markets and industries are adapting. But it is these complex systems that will adapt. To speak of going back to a pre-1900 or even pre-industrial society--and staying there for any length of time--betrays a profound lack of understanding of how far we have come in just 50 years. I daresay most of us don't really have a clue of what the standard of living was like for 95% of Americans in 1949, just as most of us don't really understand how 80% of humanity lives today. We still have all the knowledge on how to bridge those 50 years, and most of it doesn't require a massive industrial base. Note that in all this, I am definitely not the "Pollyanna" type. I am on the record publicly as estimating that the US impact will be a 7 on the 0..10 WDCY2K scale (see www.wdcy2k.org/survey). I suspect it may be less than that, though reports of recent days (e.g., the December 14th issue of BusinessWeek) don't give me much reason to hope. On the other hand, I have seen nothing to indicate that it will be an 8 or higher, much less the end of the human species. But even if we are looking at an 8 or a 9 on the scale, I believe that the speed at which we would respond and adapt would stun even us. The world might be forever changed, but that's happened several times this century already (WWI, WWII, fall of the Soviet Union, to name a few), and there's no reason why it won't keep happening. For us, two to four years of economic contraction may appear to be a disaster or even the end of the world; for much of the world, it's just a way of life. Maybe it's just time we grew up, quit whining, and got over ourselves. Life isn't easy or fair; our salaries may not always go up each year, and we may even have to go through very tough economic and political hard times as did our parents and grandparents. That's not the end of the world; it's just reality. Deal with it; we have more important work to do. ..bruce.. [For those of you interested in learning more about complex adaptive systems, I'd recommend Out of Control and New Rules for a New Economy by Kevin Kelly, as well as Hidden Order and Emergence by John Holland for starters. I have a list of another dozen or so books if you get really serious.] (c) 1998 Bruce Webster This article is published as part of Cory Hamasaki's DC Y2K Weather Report and may be reproduced under the same terms and conditions. All other rights reserved to the author. ---------- Bruce has been in software engineering and information technology for 24 years; he spent 18 months working in a Fortune 50 Y2K project, including 9 months setting up from scratch the Y2K contingency planning effort; He is co-founder and co-chair of the Washington D.C. Year 2000 Group (www.wdcy2k.org); he has testified before Congress on Y2K three times and have given private presentations to the World Bank, Congressional staff members, and the US intelligence community; Bruce's book will be published by Prentice-Hall, The Y2K Survival Guide. (Original title was The Winter of Our Disconnect, but P-H marketing overruled him.) For more details, see www.bfwa.com.] ------------------------------------- | Bruce F. Webster | | Chief Technical Officer, OSG | | Co-Chair, WDCY2K Group | | bwebster@bfwa.com | | www.osgcorp.com | | www.wdcy2k.org | | www.bfwa.com | | mobile: 202.256.1279 | | office: 972.650.2026 | ------------------------------------- Commentary -- Infomagic -- Response to my critics by y2000 @ InfoMagic.com I am surprised and somewhat saddened by some of the responses that I have received to part 2 in my series "Set Recovery On". I am, of course, delighted that so many of you have read these articles and have taken the time to respond positively, either publicly in the c.s.y2k news group or privately by eMail. (If I have not responded to your eMail please forgive me, the response has been enormous and I just haven't had the time). My disappointment stems from the fact that so many of the negative responses were more in the nature of personal attacks rather than substantive arguments against the positions I have put forward. While there is, without doubt, a great deal of emotion attached to the Y2K situation I have always tried to avoid descending to such levels, even to the point of declining to respond any further when the discussions have reached that level. Philosophically, I will do everything reasonable to avoid being robbed, but I still believe that it is better to be robbed than to be the robber. In like manner, I believe that it is better to be called names than to be the one calling those names to others. I say this not for myself but because what I am about to say might itself be interpreted as a personal attack even though it is not so intended. In spite of their personal nature, there have been many responses which, even though I consider them incorrect, are still worthy of consideration and to which I should therefore reply. To do so one by one, line by line, in the newsgroups is impractical. Therefore, I have chosen to respond to one particular person whose comments just happen to contain most of the arguments against my hypotheses. Bruce Webster is co-chair of the Washington D.C. Year 2000 Group. His comments are included in juxtaposition to my own in this outer article of Cory Hamasaki's DC Y2K Weather Report (WRP). Bruce, please believe me that this is not a personal attack. I am responding to a virtual opponent and you just happen to be the real address space within which the these pages have been dispersed. I would like to begin with your assertion that the most likely problem to stop (or limit) the airlines from flying in January will be the loss of liability insurance. (Elsewhere, Bruce also predicted that this alone might reduce airline traffic by 80% for about 3 weeks). On the face of it, this argument is relatively easy to dispose of. First, there is no aviation (FAA) regulation in the U.S. which requires an airline (or any other aircraft operator) to even have liability insurance. The need for insurance is determined by Department of Commerce regulations and, of course, by a common sense desire to responsibly manage financial risk. It would be easy for the DoC to waive the insurance requirement if the costs rise too high, to allow the larger airlines to self-insure, or even to provide government insurance during a high risk period. There is even a historical precedent for this -- the Berlin airlift. Second, the insurance companies do not (and legally cannot) decide whether it is is safe to fly, they only decide whether or not a particular airline's normal daily operations and procedures are safe enough to insure. From memory, based on specific knowledge of the wording in my own commercial aviation insurance policies a few years back (air taxi), there may not be anything an insurance company can legally do to void the policy, or to withhold coverage, as long as the airline is operating within the FAR's and within its individual operating limitations as specified in its AOC (Air Operator Certificate). It might be argued that the insurance companies will add Y2K limitations or much higher rates next year, but it is by no means certain that they can, or will, do so. There are lawsuits under way right now to determine whether insurance companies can legally apply such exclusions. In addition, insurance companies are in business to make money. If they don't sell insurance, they don't make money and they don't please their stockholders. In short, insurance is not a major limiting factor. Finally, it may be my "lack of understanding of key business issues", but I just cannot see why insurance would be available/waived for 20% of the flights and not for the other 80%. More seriously, this discussion goes to the very heart of the pollyanna/doombrood differences of opinion. My complaint is that the pollyannas tend to fixate on single, individual problems rather than stepping back to take even a glimpse of the larger problem, with all of it's intricacies and interdependencies. Moreover, the examples they use to support their position are so often totally unrelated to Y2K and its possible/probable effects. As a case in point, consider Bruce's raising of the "repercussions of the Swissair tragedy" had that occurred on January 1, 2000. This is completely irrelevant. That accident was (apparently) caused by a fire in the in-flight entertainment system, combined with CRM (Cockpit Resource Management) failures on the part of the flight crew -- what we used to call pilot error. A date of 1/1/2000 would not change the cause, or the liability, or the size of the financial settlement to the victims' families. At worst, the law suit might be dragged out a little longer while the defendants' trial liars attempt to blame the ATC system for the accident. However, this example does lead us to another of Bruce's points. He asks for a "clear understanding" of how businesses fail and, quite rightly, asserts that most businesses fail because of a lack of cash or credit. But this is like saying that an airplane crashed because it ran out of fuel, or a patient with gunshot wounds died from loss of blood. Such a view doesn't address the reasons why the airplane ran out of fuel or why the patient was shot. These might include incorrect fuel loaded in one or more tanks; unforecast headwinds beyond reasonable expectations; improper operation of the engines; failure of a simple drain valve; etc., etc. Perhaps the pilot was the patient who was shot and the airplane just flew on autopilot until it ran out of fuel! If there is one thing I have learned it is that aircraft accidents are almost never caused by a single factor but, rather, by a combination of factors, a chain of interrelated events. The same applies to the failure of a business. The running out of cash and credit is always the result of a number of different factors. My postulate is that the one-two punch of Y2K and a depressed economy will lead to a much higher incidence of such factors which, when combined, will indeed lead to a very high rate of business failures. In the case of airlines, I firmly believe that the litany of factors leading and adding to their demise will include all of those I listed earlier -- fuel supply interruptions, problems in the ATC system and generally lowered economic activity and travel, as well as the probable collapse of the banking/financial system on which everything else depends. It will also include Bruce's insurance problems (in the form of higher rates) as well as a myriad of other factors which even I do not presume to predict in advance. Any or all of these could so easily lead to "lack of cash or credit" and the ultimate demise of individual airlines and many other businesses as well. And just to make matters worse, current worldwide trading in interest futures already in dicates that there will be a definite, predictable, credit crunch in December 1999. Apple Computer might once have been able to sustain million dollar losses for years -- but not when outside cash and credit are no longer available. This is really the crux of the problem. Bruce and the pollyannas present their case for society dealing with potential problems on the basis that "we've always been able to deal with the problems in the past", with the implied assumption that we will alw ays be able to do so in the future. Bruce specifically cites the existence of large numbers of residual bugs in all of the production software used by large corporations and, quite correctly, states that, generally, we manage to work around them. He assumes that we will be able to deal with Y2K problems in the same way. But Y2K is different. Very different. First, in spite of the intentionally optimistic numbers in my original analysis, the number of Y2K failures will be far, far higher than the number of residual failures extant in our systems today. The reason for this is the sheer volume of changes we have been making in order to achieve compliance. In my experience, I have found date references in source code every 30 or 40 lines in average programs. This is one or two orders of magnitude greater than the code which must be checked and changed for a typical maintenence project. Furthermore, we haven't just changed one system. We have made major changes to virtually every system in the world, all in the relatively same short span of time. In addition, the changes have been performed under extreme time pressures, often undertaken by less qualified personnel, without adequate testing facilities (time machines) and with a far less competent level of project management and direction than I think appropriate to the size and importance of the problem. Second, the character of Y2K problems is different from the simple problems about which Bruce and others are talking. Typically, Y2K is a "boundary condition" error and, as such, it is much more likely to crash entire systems instead of just introducing minor annoyances or inconveniences. For example, if the "bounds" of an array are erroneously exceeded, the usual result is that other storage is overwritten and the entire system fails in strange, unpredictable ways. In the Y2K context, we have seen this already. The first Y2K law suit was filed by a grocery company, whose system could not accept credit cards which expired in 2000. But it wasn't just an inconvenient false date error on a printed sales slip. The entire system crashed, several times a week, whenever one of these cards was read. Each time it took hours to fix the problem and to bring the system back on line. Yes, you're absolutely right, they did still deal with the problem but it cost them enough to justify a multi-million dollar lawsuit. And that was in a fully functional economy, while everything else was still working, and they could still get support from the software vendor. When they could still get cash and credit from their bank. That was just an early error, occurring in isolation. The closer we get to Y2K, the more frequently these problems will occur and, as I showed in my original analysis, even with overly optimistic numbers, eventually the combined rate of failure will become very high indeed. At some point, next year, we will run out of programmers who can take time off from the real work of Y2K in order to fix-on-failure or to manually patch and reload a database several times a week. At some date, the number of failures will reach a point that they cannot all be addressed immediately and together. Many systems will have to wait, perhaps not just for an available programmer, but for the fixing of another system on which it is dependent. In fact, try and fix any system if the power is out, or if the banks can't cash the client's check, or the programmer can't get gas for his car. This is why I think the critical failure rate (resulting in business bankruptcies) will be as high as 10% to 20%, let alone the miniscule and ridiculously optimistic 1% I used in my original calculations. I have been widely criticized for these estimates and, I freely admit, I don't have specific numbers to back them up (Y2K really is different from anything we have ever encountered before and there are no applicable heuristics). All I can do is guesstimate on the basis of extensive prior experience. However, I am not alone in my prognostications. The FDIC (Federal Deposit Insurance Corporation) is advising it's member banks to expect a client bankruptcy rate of up to 15% as a result of Y2K problems, slap bang in the middle of the range I have predicted. Perhaps they, too, are closet doombrooders? Bruce's final point has to do with "complex systems" and their ability to "adapt" to changing circumstances. His assumption is that complex systems are always, or perhaps just usually, capable of adapting to changing conditions. As far as computer syste ms are concerned, I would tend to agree with him under normal conditions. System maintenance comes under this category and so would Y2K remediation -- if there was enough time. But time is the factor he is ignoring. There isn't enough of it left to fix (or adapt) enough of the systems and so they are going to fail, in very large numbers. The same applies to the real life social systems which computer systems only model. I argue that this factor alone will reduce the productive output (carrying capacity) of the global economy to such an extent that large segments of the population will indeed die, in turn leading to less ability to fix the systems and recover, to still lower carrying capacity, to still more deaths, and so on. This is what I call a devolutionary spiral. Perhaps a desolation spiral would be a better name, since it more accurately reflects the end results. Systems (computer, social or biological) cannot adapt if there is not enough time to do so before they actually and completely fail. This happened 65 million years ago when few life forms had the time to adapt to major ecological changes resulting from an asteroid impact. And, no, even I am not suggesting that Y2K will be that bad (I hope), but the principle is the same. What I am saying, without equivocation, that there is not enough time to "adapt" our computer systems to survive the century rollover. I am also saying that, after the failures, the impact on overall human activity will be so great that we will not have the strength to adapt our social systems to the new conditions. Eventually, we will adapt, but the conditions to which we adapt will be completely different to anything we can now imagine. Like it or not, Bruce, complexity theory and self-adaptive behaviour just don't scale very well to multiple, massive, parallel, systemic failures. (c) 1998 y2000@infomagic.com This article is published as part of Cory Hamasaki's DC Y2K Weather Report and may be reproduced under the same terms and conditions. All other rights reserved to the author.
-- cpr (buytexas@swbell.net), September 10, 2000
above from one of Cory's "Circle of Friends".http://www.sonnet.co.uk/muse/DCW-106.TXT
-- cpr (buytexas@swbell.net), September 10, 2000.