More "good news" for Flint from c.s.y2k:

Source: WSB-TV News Thursday evening Oct. 7 MARTA- Metro Atlanta Transit system announced that they are only 70% compliant. They also admitted that they will NOT be ready by January. There may be planned shutdowns if certain parts of the transit system need repairs. The impact is expected to be delays and shutdowns. They do not know the vulnerability of the automated systems and ticket systems. MARTA runs the Atlanta commuter rail and bus systems. History: Atlanta started a Y2K remediation project that was stopped by billing disputes, which brought the whole operation to a complete halt. I believe that we will see many similar announcements over the next 80 days. At risk: city 911 dispatch computers, city payroll, city water and sewer systems, and the city run hospital. So which city agency will announce planned shutdowns next? Best Regards,

Ed Hill Atlanta, GA Database Analyst/Programmer

-- a (a@a.a), October 08, 1999

Answers

84 days.

Y2K CANNOT BE FIXED!

-- Jack (jsprat@eld.~net), October 08, 1999.

One of my best friends was involved with the late start designing the new water AND sewage treatment facilities for Atlanta metro. His biggest fear is the unavailability of contracted parts to continue their work. He says they can finish both new systems by July IF all parts arrive in a timely manner. Atlanta could be a testy place to be. Earliest for water March, for sewage, June or July. Rusty

-- Rusty (eeford@bellsouth.net), October 08, 1999.

Atlanta would appear to be in a good deal of trouble. Interesting that with CNN based there, it is the center of world news coverage.

-- the Virginian (Atlanta@RememberShermansMarch.org), October 09, 1999.

Also,

I knew there was another good reason why the South should have freed the slaves and then fired on Ft. Sumter!

-- -- the Virginian (Atlanta@RememberShermansMarch.org), October 09, 1999.

Could anyone link ANY DOCUMENTS to your Atlanta fate, esp. the Water and Sewage issues. I'm going there for Christmas and will head back home the 26th. My wife's family lives there and the whole family will be there. It is of great importance to give the D.G.I. family members some insight via E-mail real soon. It may also keep me from having to go since I've insisted and argued until I was blue and still dragged into the holiday plan. Even though it was my turn to compromise, she cannot understand why I'm against the trip so much. Please HEEEELP.

-- Jaded about this one (don't wanna go@atlanta.ga), October 09, 1999.

MARTA sets halt for Y2K By Ron Martz, The Atlanta Journal-ConstitutionMore from Atlanta: At 11:55 p.m. on Dec. 31, the Y2K computer scare will bring all MARTA trains to a standstill.

But the stoppage will not be caused by any computer problems, MARTA officials said Thursday. The trains will be shut down intentionally for a few minutes, "just in case."

"Like everybody else, we just want to make sure that we've covered ourselves. We've run a lot of tests and we're more than confident that there won't be any (unintentional) disruption in service," MARTA spokesman Bobby Harper said.

The transit agency has not decided exactly how long the trains will be stopped, Harper said, but it likely will be no more than 10-15 minutes. And the trains will be stopped in stations, not between them.

Harper said MARTA also is considering stopping the trains at 12:55 a.m. for another 10-15 minutes to make sure there is not a problem with a balky computer somewhere running on Central Standard Time. Also under consideration is the possibility of shutting down station elevators and escalators.

"The only thing that would shut us down is if we lost electricity, because we are completely Y2K compliant," Harper said.

Once the intentional Y2K halts are out of the way, Harper said, MARTA's trains and buses will run throughout the night.

Atlanta is the latest major city that has decided to shut down its rail service briefly as calendars and clocks flip to 2000.

French officials say they will stop all trains and subway cars in the country between 11:55 p.m. and 12:15 a.m. in order to "respond to the anxiety" of passengers, according to a statement issued earlier this week by the national train company.

The Chicago Transit Authority also will pause its rail service briefly from just before midnight until just after the new year begins.

Of 550 federally funded operators surveyed by the Federal Transit Authority, only four indicated they were not Y2K compliant. All of them are in Puerto Rico.

But some transit officials worry that the confluence of potential Y2K computer problems and massive New Year's celebrations for the Year 2000, mistakenly thought by many to be the start of a new millennium, could produce problems for the tens of thousands of revelers being encouraged to use mass transit.

Harper said about 500,000 riders used MARTA last New Year's and at least an equal number is expected this year.

In Washington last month, a computer that monitors all the trains in that metro system locked up and delayed morning commuters more than 50 minutes.

Atlanta commuters will be warned of the planned New Year's Eve stoppages in an advertising campaign that will begin later this fall.

-- the Virginian (WhereIsTheTruth@Atlanta.com), October 09, 1999.

Okay. So far, we have airline decrease/stoppage of flights, rail service stopped, and chemical plant slowdown/shutdowns. How will that sudden loss of load affect the utilities?

-- Rachel Gibson (rgibson@hotmail.com), October 09, 1999.

Interesting. I know that some embedded devices take time samples at known intervals. Some very small percentage of these use the entire date (including the year) for this purpose. And such devices can potentially become confused when two samples are taken, the first in this century and the next in the next century. But this problem arises only between those special samples, and nowhen else.

However, all such devices I'm familiar with also have cold-init code that must necessarily assume no 'prior' sample and start with known default value and continue from there. For this entire class of device, shutting down before turnover and restarting afterward dodges that one rogue sample, and hence dodges the problem.

Note that for this to work, the time being kept by the device must be fairly accurate, meaning it requires a source of updating from somewhere. Otherwise you get clock drift. And this must be fairly often, since those clocks can drift from real-world time by an hour a year or more. I'm sure the Atlanta engineers are aware of this.

I suppose most of these short-term (10-minute) shutdowns we read about are intended to address that sampling issue. But from this material, I can't guess whether they *know* this is what's happening, or are just crossing their fingers. In either case, (assuming accurate clocks) they're guarding against some wildcard misbehavior at the rollover, and planning to restart under controlled conditions.

Presumably the utilities can handle the reduction in load when it's announced and scheduled in advance, and doesn't come as a total surprise.

-- Flint (flintc@mindspring.com), October 09, 1999.

Rachel: Much of our electricity is produced in dedicated coal- burning power plants. Lots of coal moves by rail. Less coal, less power produced.

www.y2ksafeminnesota.com

-- (The Real) MinnesotaSmith (y2ksafeminnesota@hotmail.com), October 09, 1999.

Flint and MS, I was approaching the topic in a global sense and not within the narrow confines of Atlanta, although this city appears to be a prime example of many things that could go wrong all at once (see other threads discussing its water supply and its airport, for example).

Numerous discussions have also covered types of power generation and distribution. Without going into those specifics, and without touching upon those high-load usage companies that stop operating unexpectedly, I suspect that the utilities may not be fully informed of the planned changes/stoppages, especially of last-minute ones as we near rollover.

My utility (power provider) stated last fall that it is most concerned about loss of load. This same utility is currently laying-in large stocks of food for its employees.

Flint, I appreciate your discussion of time measurements; for the non-experts amongst us, why would the systems not use a time measurement that came from just before rollover? Or is that too technical, too much an unknown?

-- Rachel Gibson (rgibson@hotmail.com), October 09, 1999.

Rachel: Do you like to mudwrestle?

-- King of Spain (madrid@aol.cum), October 09, 1999.

Rachel:

I'm not sure I understand your question. Devices that are continuously sampling the time determine the interval by subtracting prior time from current time. If the year is used in the sample and the year is 2 digits, this can cause the code to misinterpret the length of that one interval during which rollover happened according to the time source being used.

Usually, depowering the device and "rebooting" it causes it to execute power-up initialization code. This code "knows" that there is no "prior" time sample. Typically, this code contains a one-time flag telling it that this is the very first sample, so *don't* calculate any interval from it -- there isn't one. From the viewpoint of the device, time starts on that first sample, so there's no interval until the second sample.

Therefore, shutting it down before rollover and restarting after rollover skips the troublesome sample, and all is fine. And note that this only applies to such a device, which is common but by no means universal. If the embedded system does this at a high level, it might actually have prior samples in some kind of filing system. Such historical data must be deleted for a true start-up condition to happen.

-- Flint (flintc@mindspring.com), October 09, 1999.

Flint, you lost me. "Interesting. I know that some embedded devices take time samples at known intervals." (from your first response) Are you referring to devices used by utilities in relation to loss of load, to devices used by trains, planes and chemical plants, or to both? Or to neither?

KOS Sigh This is the second time you've asked me. Refer to the first time for my response.

-- Rachel Gibson (rgibson@hotmail.com), October 09, 1999.