To Flint - Re: IM report

greenspun.com : LUSENET : TimeBomb 2000 (Y2000) : One Thread

Flint,

Since the 10% failure rate post is about to slide off the edge of the world, I thought I would bring this back to the top. I think this topic deserves a bit more attention, and I hope the rest of the forum agrees.

First, thanks for your well thought out and civil response. Second, pardon the delay on my part. I just got finished stacking 4 cords of firewood!!

We agree in general on the fact that the IM report is of dubious value, I believe. We just arrived at that conclusion via different routes. BTW did you visit their website? Very professional! Quote "International Monitoring provides conultation and resource allocation advice...." *spell checker ON please!!*

While they undoubtedly stand to benefit/profit from Y2K, I still think the main issue here is that you cannot predict anything worthwhile using information coming from countries that are just in the "forming a committee" stage. That's all I'm saying.

The more important point, in my mind anyway, is concerning the "behind schedule" business. Various organizations that attempted to create standards with regard to Y2K remediation were all in agreement that a full year of testing was necessary. Nobody argued that point. Wellsir, here we are. You can count on one hand the number of companies/agencies that made that deadline. How far can we push the envelope? Certainly it would depend to some extent on the size and complexity of the systems involved in each outfit, except that in this food chain, how can the big boys test end-to-end if the little guys - suppliers, vendors, contractors - are waiting to fix on failure? Or pushing the envelope to November or December? How can any of the Bigs claim compliance at all knowing this?

What we are looking at, I'm afraid, is what Cory Hamasaki described some time back, where eventually the number of glitches/failures exceed the number of IT personnel available to fix them all. If everyone had met that December 1999 deadline, and everybody was now in testing end-to-end, I wouldn't have been stacking firewood all day!

So, are we looking at the end of the world? Naw, probably not. But we are looking at the end of a lot of people's own little worlds. A lifestyle adjustment of major proportions for most people, IMHO. And anyone who thinks it will all be fixed in a matter of 2 weeks is smokin' some pretty good stuff.

-- Don (dwegner@cheyenneweb.com), July 01, 1999

Answers

I can't imagine there to be any arguement with this, Don. About the only position that 'could' be taken might be, "It's not over, 'till it's over". This has never been much of an arguement as far as I'm concerned. Not if one carefully examines the complexity of the Y2K beast, and then looks beyond our national boundries. *JULY 1999*

BTW Don....we had a yearly tradition of spending the 4th of July up in Hog Park. It snowed 3" on us up there in '94 (?). We never went without our insulated coveralls, ANY year! I couldn't even begin to tell you how VERY much we miss that trip. That firewood is a tough job, but absolutely required! Good luck, lots of prayers aimed in the direction of your neck of the woods!!!

-- Will continue (farming@home.com), July 01, 1999.


Don:

I mostly agree with you. I've forgotten what I wrote earlier about these schedules, so forgive me if I repeat myself any.

1) What we're seeing is obviously messy. An adequate testing period depends on the size of the testing task. For a one-PC small business, this might be no more than one day. For a large organization, this might be well over a year, depending on what they needed to do and many other factors including remediation skill and resources applied, nature of the task, etc.

2) Bear in mind that the one-year-for-testing policy was promulgated only by a tiny minority of businesses. Yes, Gary North picked it up and has been chanting it so monotonously that we forget that it was never a position taken by the vast majority of organizations.

3) There is (at least to some people) an implication that organizations draw a line that basically says, No testing of *anything* before remediation of *everything*. This is universally false. Probably no two organizations schedule assessment, classification (critical/non critical), remediation, unit testing, system testing, and inter-system testing the same way. Their entire compliance effort is atomized into numerous elements, each with its own appropriate schedule and timeline.

4) Many of these schedules were unrealistic. I won't even try to guess who in any organization actually believed otherwise. There is a philosophy pretty widely held that greater speed is achieved by setting an aggressive target. Otherwise, Parkinson's law operates -- work expands to fill the time available.

5) Internally, a remediation project has many deadlines and milestone dates. These are often impractical to summarize. Usually there is a critical path, and many (probably most) of these critical paths are being lengthened. No question many firms underestimated the size of the task facing them.

6) As you point out, economy-wide end-to-end testing is a pipe dream. Even if *everyone* felt fully compliant, any such testing would still be beyond our ability to define, much less organize. Hell, it might be years before more than 3/4 of the actual remediated code is tested in real life. For this reason, I agree with IEEE that y2k is a long- term problem.

7) Within each organization (integrating all their various remediation projects together for convenience) I picture a "compliance curve", such that each N hours of effort produces less results than the previous N hours. Total compliance cannot be reached. The number and severity of experienced date handling bugs depends on where the organization is on that curve when the bugs bite. The closer they are to "full" compliance, the lower the number and severity of bugs they must handle later.

The same curve applies once bugs start biting. Most can be handled within a week or two, and some will drag on and on. If the number and severity of the bugs that drag on is sufficiently large, the viability of the business itself is in jeopardy.

What we have is millions of both of these curves, no two alike anywhere. NOW we need to determine how our daily lives will be affected. It seems clear to me that we can't even begin to quantify this. It will be different everywhere and for everyone. Like shooting into a milling crowd -- we know some people will be hurt, but cannot predict who, or how badly. At best, we can try for a ballpark estimate of how many bullets will be fired.

My own feeling is that the large majority of the crowd will still be standing after the shooting (slowly) dies down. But the whole experience won't be pleasant for anyone, and extremely unpleasant for an unlucky minority.

I'm also convinced that no matter *how* early everyone had started (or how hard they worked), the number of bullets would never be zero. Out entire remediation effort is intended to reduce this bullet count to a "manageable" level. But some will be hit. No doubt about it.

-- Flint (flintc@mindspring.com), July 01, 1999.


**Will** Thanks for the thoughts and prayers. Yes sir, that's a great area. I've been snowed upon many a summer day up around Medicine Bow Peak, fishing in the Gap Lakes. You have family up here? Feel free to e- mail me privately.

**Flint** Now, I must admit, I am somewhat puzzled. Around here, I am considered a fearmongering doom and gloomer. On this forum, you are considered a polly. Looks like our views are not all that far apart. Go figure. Maybe I should seek counselling to help sort it out ;)

-- Don (dwegner@cheyenneweb.com), July 01, 1999.


Moderation questions? read the FAQ