http://azstarnet.com/~nuu/Other_DOCS/charles_perrow_2.htm

Fascinating reading. "Electrical Power" is the lengthy concluding section. Rick Cowles is quoted frequently.

Mercifully, I spare the uninitiated from the terrors of excerptation.

-- Anonymous, June 18, 1999

Answers

Lane,

This is a fascinating report. Three groups, pessimist, skeptic, optimist. Now I know why I'm a pessimist, with regard to Y2k. Well, actually, I have know this for quite a while now but never could explain it in quite this detail. It's a long read, for over the weekend. Print it out even. I think one of our regular posters, Flint, is a skeptic. What say you, Flint? Does this essay make any sense to you? Do you think it discusses the matter in enough detail?

-- Anonymous, June 18, 1999

Go get the book.

(I'm emailing Charles to see if the updated version has been published)

Get it even if it doesn't have the Y2K update in it.

Read the book.

Then, and only then, will I listen to you speak about "there may be isolated failures."

Critt (who's not buying this Bump-In-The-Road poppy-cock)

-- Anonymous, June 19, 1999

A thoughtful and interesting article. But it has a slant that makes me a little dubious. Without going into blow-by-blow detail, here are some points that bother me:

1) The article talks about "failure" simplistically, implicitly assuming that the economy resembles a stick rather than a reed. I don't think past experience, however incompletely applicable, supports this implication. The possibility (yes, it's only a possibility) that systems will degrade rather than break outright isn't given credit here. If, say, a manufacturing plant suffers N% reduced output for X period of time, just how large do N and X need to be before we can say that plant has "failed"? Trying to estimate the *degree* to which we'll suffer broad inefficiency and slowdown, is not well served by speaking exclusively in terms of "failure", yes or no.

2) When writing of "complex interactive systems", I notice the careful avoidence of the word "adaptive", both explicitly and implicitly. What ends up being pictured is a static system of some description. To the extent that this static system is like a web rather than a chain, alternative methods, paths, and approaches either exist or they do not. Yet all of our systems are adaptive to some extent. We fix what's broken, and/or we find alternatives/ and/or we create alternatives where they didn't exist before because they didn't need to exist. For almost all of our processes, there are or can be suboptimal substitutes, at every level. We don't use (or haven't created) these substitutes because we haven't needed to and they are less cost-effective.

Along these lines, y2k isn't a very good measure of the robustness of our systems, because there is no control to this experiment. As a thought experiment, if our systems are fragile but y2k bugs are mild and don't cause large or cascading problems, we can't therefore conclude that our systems are robust. Conversely, our systems may be highly resiliant and still be overwhelmed if y2k bugs are sufficiently widespread and severe. To measure either, we must hold the other constant, and we can't do this.

3) While I'm not expert in electric utilities, I do know all about the real time clock used in PC's, and I know that the statement here summarizing Dallas Semiconductor's assessment is simply false. This makes me nervous about some of the rest of the article. The real time clock used in almost every PC uses 2-digit years. The use of 2-digit years by itself does *not* constitute noncompliance in any way. Noncompliance stems from *misuse* of 2-digit years by the system software. If these years are properly windowed, there is no noncompliance at all.

Now, let's say the system BIOS ignores or bungles the windowing. Yes, this is noncompliance. So after rollover, you examine the date and reset it if it's wrong. One time. Problem solved, with very rare exceptions. Yes, there might be problems with application software, but replacing the PC won't help with these a bit. Summarizing a problem found largly on 486 and older units, which is trivial to fix, as "all PCs are noncompliant" implies either a slant or technical ignorance. Either one is cause for pause when reading the rest of this article.

4) In a similar vein, the original GM and Chrysler evaluations are mentioned, but no mention of the subsequent efforts and successful testing of those efforts. The article pays some lip service to the notion that not everything that can go wrong, will go wrong. But it spends a lot of time talking about a great deal that can go wrong, and no time talking about what has been remediated, or what testing has shown doesn't actually go wrong in practice.

None of this should be taken to mean that I don't expect problems. I expect many. Which systems will be overwhelmed by either number or severity of bugs, and what will be done to address them, is highly contingent on factors beyond our ability to estimate. A lot depends on implementation details, remediation skills, and experience of the troubleshooters later. There will be many surprises, most of them unpleasant but not all of them. If y2k is a bump in the road, it's a big one and millions of vehicles will encounter it. Some will get past it with few problems, and some wheels will fall off.

-- Anonymous, June 19, 1999

In all of my reading of this and other threads regarding the grid, a number of things strike home:

1. Y2K will break things. But that is nothing new, they are breaking every day. Just that Y2K may break many more than have been broken before.

2. Once you get to this point, it all becomes a question of the attitudes of each of us individually and collectively whether this will be a "failure"

3. This is wasted argument. Who cares whether you call it a failure or a success. The real question is whether the society keeps its collective head regardless what happens. Its not an issue whether the electrical grid holds but whether the psychological grid holds.

4. In this regard, my hat is off to the can-do attitude and optimism of the engineers who tell us they will be ready. Yes, I know that even if they are 100% ready, they will only be 85% ready. And I know that Gary North predicts cascading So what? It's their determination to be ready and to beat this thing that will win the day.

5. Why this need to be right on this issue? I suspect the essence of it is our desire for certainty before we start running the risk of looking like Y2K Whackos in front of our friends and neighbors. Everybody knows that if the lights go out, our honor is saved. Hence the interest in this general discussion (and perhaps a very subtle but macabre desire to believe the bad news over the good.) FORGET THE CERTAINTY. THE LIGHTS WILL PROBABLY STAY ON, AND THOSE OF US WHO PREPARE WILL LOOK LIKE DAMN FOOLS. ACCEPT IT. (And understand that if you are right, if there are significant problems, your present anxiety is the price of leadership.)

So let's stop worrying about what the "right answer" is here. Let's learn what we can, make our preparations, and be ready to do whatever we can to turn a potential catastrophe into another triumph of the American Spirit.

-- Anonymous, June 19, 1999

Bill,

That was a very good statement from you. I appreciate the honesty that rings thru it all. Some of your thoughts I agree with, some I disagree with, some I will have to think about more. One thing you said though, which strikes me as 100% accurate, and which I relate to deeply is this: "And understand that if you are right, if there are significant problems, your present anxiety is the price of leadership." It's lonely at the top, some say, and I can feel that now, as I'm sure other leaders here can as well.

-- Anonymous, June 19, 1999

Flint - nice to see you posting here!

-- Anonymous, June 20, 1999