"More Guesses, The Days After 2000-01-01" (Dick Mills - April 9, 1999

[snip]

In 1588, while awaiting news of the coming of the Spanish Armada, Sir Francis Drake wrote, "The advantage of time and place in all practical actions is half a victory; which being lost is irrecoverable". He could have been talking about the deployment of electric utility human resources in 2000.

[/snip]

http://www.y2ktimebomb.com/PP/RC/dm9914.htm#Conclusion

~C~

-- Critt Jarvis (middleground@critt.com), April 10, 1999

Answers

I thought that forum participants may be interested in reading email exchanges I've had with Dick Mills with regard to this article.

Dick:

I read your article, "More Guesses", and while I think you provide a good BEGINNING, I still think your estimate of a blackout, and manpower-requirements-to-operate-in-manual-mode are too optimistic.

Because Y2k is systemic in nature, you must apply the Drake methodology to take into account ALL relevant factors which have a reasonable chance of contributing to your final estimate. Let me give you some factors that you left out that, in my opinion, should significantly contribute to your "blackout" and "manual mode manpower" estimate:

*Include a similar Drake expression for power transmission. *Include a similar Drake expression for power distribution. *Since power plants require SOME functional elements of our telecommunications system to work properly in order to deliver power, include a Drake expression for at least a partially functional telecom system. *If, as you say, operating problems will persist for months, then a reliable banking system will be essential. You may glibly dismiss this assertion as knit-picking, but until proven otherwise, the probability of a reliable banking system must be included. * Probability that an electrical facility (power generation, transmission, or distribution) will put together an effective "manual mode" contingency PLAN. * Probability that such a plan will be ready in time and tested adequately. *Probability that the effects of Y2k will be moderate enough to allow ALL employees (both manual crew and normal operating crew) to get to work, and, as a separate probability, that local materiel and service support will be available. Since your "half life" is on the order of 4 months, the probability of fuel delivery will come into question. Therefore, you must include the following: Probability of coal and crude oil production will be adequate (because foreign crude availability is in question, it should be given a separate probability). Probability that pipelines and rail transport will operate adequately. Since there is a significant chance that "outside forces" (such as cyber-terrorism, or multiple provocative hostile acts occurring around the world) will take advantage of Y2k to create disruption, you need to take that into account by assigning a suitable probability that foreign delivery of oil/gas, and spare parts could be significantly reduced or even eliminated. What about a whole host of "indirect" factors such as, *Probability that any electrical facility will be deemed "unsafe" and not allowed to restart by the relevant governmental agency. (this estimate should include the probability that local fire fighting equipment, for example, is hobbled by Y2k events). *Probability that water supplies are unavailable. *Probability that sewage systems are crippled by Y2k.

I'm sure a professional such as yourself would be able to add many more items to my list. Incidentally, if you don't think what I'm saying makes sense please explain why in a future article.

Dr. Roger Altman

xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

Roger,

I can think of thousands and thousands of risks. That makes ALL much too big. The readers would get weary if it were too long. Case in point, you seem to have not read to the end of the article yourself, nor read the comments I've made on some of these subjects in past articles in the series.

"ALL reasonable" is more manageable. I've given what I consider the the most reasonable ones and said that "these are the lions share" If you have other calibrations of what's reasonable, then you're perfectly free to write your own equation an to publish it together with your qualifications and your reasoning.

Until then, since I'm the only one who's stuck his neck out publicly in this manner to explain my reasoning and my quantitative judgements, I'll be the sole judge of reasonableness and set my own boundaries and my own schedule.

By my judgement, I've already given the most reasonable factors.

xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

Dick:

I fully sympathize with your concern that readability would suffer if you followed my "systemic" approach. However, a back of the envelope estimate of the probabilities corresponding to these additional risk factors may convince you that it's worth the trouble.

Assuming that it is, setting up the equations to show the additional risk factors is all that would be necessary in the actual article. And since these additional equations would also be based on Drake's logic, your readers wouldn't be asked to understand any more than they do already. A simple table could be constructed to show your probability estimates; No big deal especially if the results would come a lot closer to reality.

Give it a try, my seat of the pants estimate paints a much more gloomy picture of what we may be in for. If I'm right, don't you owe it to the public to give them your best effort even if it makes us all feel uncomfortable.

Sincerely yours,

Dr. Roger Altman

-- Roger Altman (RogAltman@AOL.com), April 10, 1999.

Mr. Yourdon comments on April 1 rollover

asked in the TimeBomb 2000 (Y2000) Q&A Forum ---------- Mr. Yourdon writes a weekly Y2K column for some outfit called Cutter Consortium. He wrote one of these columns about the April 1 rollover, and a friend passed it along to me. Not sure if it's kosher to post it here on his forum, but I figure that what goes around comes around...

Lurker

*******

Welcome to the Y2000 E-Mail Advisor, a weekly electronic briefing from Ed Yourdon, Director of the Cutter Consortium's Y2000 Advisory Service.

THE 1 APRIL FISCAL YEAR ROLLOVER

It's now been a week since New York, Canada, and Japan celebrated the beginning of their new fiscal year; as I write this week's column, the British government is going through the same process. Thus far, it appears that there have been no significant Y2000 problems, and Y2000 pundits are now trying to decide what it all means.

Unfortunately, much of the discussion has degenerated into petty bickering between the Y2000 optimists and pessimists (or, to use the more disparaging terms, pollyannas and doomsayers). The pessimists may have been hoping that dramatic Y2000 problems would emerge in Ottawa, Tokyo, London, or Albany in order to validate their predictions about Y2000, or to at least raise the level of awareness about potential problems looming ahead in the next few months. And the optimists may have been hoping that the lack of 1 April problems would prove, once and for all, that the pessimists' predictions were grossly exaggerated.

I found myself dragged into the fray when a participant on the comp.software.year-2000 Internet forum reminded his fellow participants shortly after 1 April that I had written a grumpy article last July (available at http://www.yourdon.com/articles/y2kClintonspeech.html) in which I said "On April 1, 1999, we will all watch anxiously as the governments of Japan and Canada, as well as the state of New York, begin their 1999-2000 fiscal year; at that moment, the speculation about Y2K will end, and we will have tangible evidence of whether governmental computer systems work or not." It was suggested that I should "put up or shut up": unless there was tangible evidence of Y2000 problems resulting from the fiscal year rollovers, I should acknowledge that Y2000 isn't going to be such a bad problem after all.

Well, maybe it won't be -- and if that's the case, we should all rejoice. After all, we're all on the same side, in terms of our desires for a best-case Y2000 outcome, even if our current assessments of the situation may differ. I'm happy to agree with the notion that the fiscal year rollover success enjoyed by New York, Canada, and Japan -- coupled with the similar experiences enjoyed by the Eurocurrency projects, as well as the absence of catastrophic rollover problems on 1 January 1999 -- means that there is less reason than before to assume that Y2000 will lead to TEOTWAWKI (the end of the world as we know it). But I never believed in the TEOTWAWKI scenario in the first place, and I still think there's an enormous amount of potential for moderate disruptions between the "bump in the road" scenario predicted by the optimists, and the apocalyptic TEOTWAWKI scenario articulated by the doomsayers.

Looking back on the words I wrote last July, I do regret having said "at that moment, the speculation about Y2K will end," because it has become evident that there is still much to speculate about. The most significant aspect of speculation involves embedded systems: regardless of whatever problems the various government authorities might have had with their financial computing systems on 1 April or their Eurocurrency systems back in January, none of it involves the embedded systems. We still don't know how that part of the Y2000 story will unfold, and we continue to be whipsawed between optimism and pessimism as we see each new report indicating that embedded systems failures are worse than -- or, according to the next report, less serious than -- what we previously believed. Y2000 optimists and pessimists will continue arguing passionately about the presence or absence of life-threatening embedded system failures until much later this year -- indeed, possibly right up to midnight on New Year's Eve.

It's also fairly clear that a fiscal-year rollover phenomenon only concerns those computer systems that are aware of the concept of a fiscal year, and that make use of that concept in their decision- making and/or calculations. Thus, there are almost certain to be a wide range of computing systems within a government organization that are date-sensitive, and potentially noncompliant, but which are completely unaffected by the fact that the organizational entity has moved into a new fiscal year that extends from 1 April 1999 to 31 March 2000.

But let's put these two categories of systems aside for the moment, and focus on the heart of the debate: what can we conclude from the apparent fact that FY-sensitive computing systems in the Canadian, Japanese, and New York State governments appear to have survived the 1999-2000 rollover event?

The optimistic interpretation is fairly obvious: it means that appropriate remediation efforts on those systems were indeed finished on time, rather than falling behind schedule; and it means that the testing efforts were sufficiently thorough and comprehensive that no "show-stopper" bugs have appeared. And if that interpretation is correct, one might logically conclude that the government agencies will extend their success to the rest of their systems, and thus succeed in remediating and testing all of the other systems that would otherwise fail on 1 January 2000.

But this gets back to the issue of what I referred to as "tangible evidence" in my 1998 article. How do we know that New York State's mission-critical systems survived the fiscal-year rollover? Well, we have the governor's word for it: on 1 April, New York Governor George Pataki issued a press release (available at http://www.state.ny.us/governor/press/year99/april1_1_99.htm), which announced that "New York State's "mission critical" computer systems -- such as the state's payroll, general accounting, and tax systems -- that are dependent on the state's fiscal year have been remediated, tested, and are in production." Alas, we have become such a cynical nation that we're not entirely sure if we can trust such public statements; after all, we live in a nation where the meaning of truth is sometimes determined by what the meaning of "is" is. Thus, the Y2000 pessimists -- along with at least a few computer- illiterate cynics -- may seek their own evidence that the state's computer systems are working properly.

But it's likely that many computer problems that might have occurred on 1 April were simple enough that they could be fixed by computer programmers within a matter of hours, without the public ever becoming aware of the problem. True, this might have cost the taxpayers some money, and it may have caused an incremental, temporary decrease in the efficiency and productivity of the state government. But the reality is that if the problem is small enough -- or, to put it more cynically, if the problem is capable of being hidden and covered up -- then it doesn't qualify as "tangible evidence." If the citizens and taxpayers don't see it, it doesn't exist.

That seems to have been the experience with the Eurocurrency situation, and the FY-99 rollover experience. That's not to say that either of these two "trigger dates" were completely problem-free; a Eurocurrency problem with one of the French banks led to riots in Marseilles by welfare recipients who had not received their checks by mid-January 1999 (see http://www.telegraph.co.uk/et?ac=000144088531503&rtmo=3umxxY3M&atmo=rN 00bsGq&P4_FOLLOW_ON=/99/1/7/wrio07.html&pg=/et/99/1/7/wrio07.html

for details); and rumors abound of European banking difficulties with bungled deposits and funds transfers. And though the FY-99 rollover didn't lead to any catastrophes, there is no shortage of minor problems that have been reported (see http://www.greenspun.com/bboard/q-and-a-fetch-msg.tcl?msg_id=000RSP for one such list).

But thus far, we have NO evidence of FY-00 rollover problems in Canada, New York, Japan, or England. It's worth remembering, though, that many of these problems -- if they exist at all -- may not show up until several "cycles" of processing have occurred, e.g., when monthly or quarterly reports have to be generated, or when several payrolls have been run. The cumulative effect of software errors -- sometimes referred to as the "Jo Anne Effect" problem among Y2000 cognoscenti -- may not be visible until numerous database records have been clobbered, or numerous end users have been affected.

And that will be the ultimate test: if FY rollover problems cause New York, England, or Canada to bungle the payment of pension checks to a hundred thousand retired civil-service workers, it won't be possible to hide the problem. We've already seen one such problem, though it's now attributed to human error rather than a Y2000 software bug: the recent fiasco that led the New Jersey state government to erroneously credit thousands of food-stamp recipients with an additional payment.

For whatever it's worth, the public acknowledgments of the 1 April situation by New York and Canadian Y2000 managers was cautious. A 2 April article in the *Los Angeles Times* (see http://www.latimes.com/excite/990402/t000029233.html) quotes Jim Bimson of the Canadian Year 2000 office as saying, "So far it's been a nonevent. We haven't heard anything today, but I'm not that surprised since we really have to wait a while for some transactions to occur. Most of the computers are still working in the last fiscal year."

Meanwhile, it's also important to remember that Canada, Japan, England, and New York State are not finished with the systems that MUST work on 1 January 2000. A quick look at the "top 40" mission- critical systems in New York State (visible at http://www.irm.state.ny.us/yr2000/top45lst.htm, which shows the status as of January 1999) indicates that roughly half are not yet compliant.

Whether you're an optimist or a pessimist, it's important to remember that we still have several months before we'll really know how Y2000 will turn out. If you think of it as a 10-round boxing match, the optimists can claim to have won the first three rounds; but there are still several rounds to go.

-- 32356 (3@23.56), April 10, 1999.