This post on TB2000 speaks of another potential problem or two that we are watching. To see the entire thread, use the link below.

http://hv.greenspun.com/bboard/q-and-a-fetch-msg.tcl?msg_id=002HFE

"...Our team saw two embedded failures this week, even with 75%-80%of the plant shutdown for a normal winter inventory adjustment. Why did they fail on 1-4-2000 and 1-6-2000 and not on 1-1-2000? I think there are a number of reasonable answers.

1. A large % of the embedded chips that are sensitive to roll-over failures were shut down anywhere from a day to a week before roll-over. They are now on-line and are truly timebombs waiting to go off. That is one of the reasons for the increasing level of failures we are seeing now.

2. Many embedded chips in control devices (PLC's, loops etc.), perform sequencing type events or process sequencing type batch jobs shutdown a process, up-date a loop, change op modes etc.) but they do this only at specific times. An embedded system in your process may look like it's OK right now, but when it's called upon to do it's thing it can't, or gives the wrong info. because its really lost it's mind. (program). These sequencing type events may happen hourly, daily weekly monthly or yearly, or any combination therof.

3. The failures I've mentioned above were at the low to mid level with respect to seriousness. No lives were in danger although it compromised a process and will cost the company $60,000 to fix. This news you will never read about in the Wall Street Journal or see on the evening news, why.... because with the current rulings in Congress, the Insurance and potential liability issues, any company would be foolish to release such information.

If and when it does get out it will be... "Not Y2K related".... count on it!"

greenem31@aol.com, January 09, 2000

-- Jennifer Bunker (Salt Lake City, Utah) (Jen@bunkergroup.com), January 09, 2000

Answers

Folks,

First I'd like to direct your attention to the website I've put-up to track the Y2K events. It relies heavily upon the GICC reports, and gives credit accordingly.

Y2K International Watch) || Address: http://www.iol15.com/coggeshall/Y2K-International-Watch/Update.htm

Second, the primary reason for my reply is to address the "embedded chips" issue.

As an Aerospace Mechanical Engineer, with heavy Petrochemical experience, I was indeed "shocked" that embedded chips didn't crash most of the US Electric Grid and Refineries the first week in January.

Indeed, I predicted such, but was entirely wrong.

So it's with a "great" deal of hestitation that I fundamentally agree with the (above) post.

Y2K failures from embedded chips SHOULD STILL BE A GRAVE THREAT.

I'd be dishonest (PLEASE don't take this the WRONG way) if sub- consciously I'm still "looking" for Y2K vindication, and embedded chips is my "twisted" psychological ticket.

On the other hand, after years of rummaging about Petrochemical Control Systems (and designing a few of my own), something tells me that "failures must come".

I haven't heard ANY reliable reports that the US Petrochemical folks spent the "billions" as per the Banks, when I always felt the most spectular failures would be Refineries and Power Plants and Offshore Oil Ports and Water Plants going up in flames, as a side-effect of faulty cooling water sensors, leak detection devices, valves, etc.

So, that's my take, for what it's worth.

Final point about embedded chips, however ...

They usually control "cycles" in the above facilities. Control Cycles can last minutes, days or even weeks, depending on what they control. It's the "longer" cycles that may yet vindicate the Jim Lords and Gary Norths of this world (I'm counting myself in with them).

But then again, maybe we're all "just full of #####".

Best Regards

Douglas A. Coggeshall

-- Douglas A. Coggeshall (coggeshall@iol15.com), January 09, 2000.

Isn't it possible that the embedded chips don't keep accurate time? I know that I constantly have to change my system clock on my computer -- it's always running slow.

-- Michael Bennie (mbennie@lightspeed.net), January 10, 2000.

It is very likely that the embedded chips don't care what the date is. Y2K Failures on occur when a date calculation is being attempted. The reason the banks spent billions on the Y2K bug was because their systems perform millions of date calculations for Interest Accruals etc. If there is no date processing being performed, there is nothing to worry about.

-- Lee Barrentine (Wrknman042Legacy@aol.com), January 13, 2000.

The last post regarding "it won't fail unless it involves date processing" was disputed by the NIST conference of embedded systems experts in November 1999. Your mileage may vary.

-- Bud Hamilton (budham@hotmail.com), January 14, 2000.