To fix a system requires a 'real time' test, and 99% of real time tests to date have failed. So, since few larger utilities, like banks, electricity, water, phones etc haven't even been real-time tested, they too will fail upon rollover. Its over folks. That is the reality.

-- owl (new@new.com), August 24, 1999

Answers

Owl,

Do you have any data to support your 99% "real-time" failures?

Keepon

-- keepon (vacillating@hourly.edu), August 24, 1999.

keepon, check out the refinery explosions, the mci failure, the pwer outage in chicago, the passport screwup in britain, the sewage dumpage in LA, the power outages where realtime tests have taken place. You see, the system was designed from the ground up to accommodate 2 digits. You can't convert it to 4. It can't be done at all, period. All these failures prove this. Have you heard of a single company claim geniune compliance? No, because all the realtie tests have either failed, or aren't be done (is AT&T gonna take that chance? ha.)

Owl

-- owl (new@new.com), August 24, 1999.

Actually, that is four sentences. Now I wonder, hmmm, do I take the advice of someone who can't even count the number of sentences they write? Nevermind that they probably have no experience or knowledge of any of the systems the poster mentions, and that they definatley have no knowledge of the 'real time' tests that, apparently, 99% have already failed. Instead, I'll just take these four sentences as law and go and build a bunker 50 feet under the ground and wait for the year 2035 when everything is back to being peaches and cream, because this owl person says so.

ya, right.

(four sentences, one fragement)

-- (owl@is a fruitcake. org), August 24, 1999.

that post wasn't meant for dorks like yourself. go lurk in the closet, where you belong doofus.

owl

-- owl (new@new.com), August 24, 1999.

Ah... now I'm convinced you know what your talking about!!

Thanks for clarifying that

-- (owl@is a weak minded. fruitcake), August 24, 1999.

I believe that the MCI foul-up had nothing to do with Y2k remediation but rather was a normal Application System upgrade within their frame relay network. You would be suprised how often this happens in the telecom world - especially when one considers how complex the AS have become...The last AS that our company developed (and that I participated in) took 1 million man-hours to develop and it was not a total re-work from the groundup. There are tests being conducted on telecom systems involving test plants and parts of the live network. Regards,

-- william holst (w_holst@hotmail.com), August 24, 1999.

owl, correct me if I am wrong, but I assume that what you are actually saying is that 99% of real-time tests fail initially -- e.g., software has been remediated, now it gets tested, and problems show up.

There is a lot of controversy as to how sufficiently even remediated Y2K code has been tested. This time last year, it seemed like the most common claim was, "We will have finished remediation by 12/31/1998, leaving a full year for testing." At the time, testing seemed like a very important thing to do, with estimates to the effect that testing should comprise 50%-70% of a Y2K project.

Its a year later, and it would appear that for the most part, its "We will have finished remediation by 9/30/1999 leaving a full three months (minus Thanksgiving/Christmas holidays) for testing."

I know that its because I am a "doomer", and suffer from some "infectious meme" or whatever, but I really do have this terrible belief that 1/1/2000 will come on schedule, and untouched or badly remediated code (including embedded chips) will execute and fail because of non-existent or inadequate testing. And I think a number of people are starting to wake up to this.

Y2K CANNOT BE FIXED!

-- Jack (jsprat@eld.net), August 24, 1999.

Correct me if I'm wrong, but isn't it true that as a whole(grid so to say) the phone "network" cannot be tested in "real time"? I heard that in order to test it like that, it would have to be taken "off line" and they won't/can't do that--for obvious reasons.

-- CygnusXI (noburnt@toast.net), August 24, 1999.

I don't know what percent of real time tests have been performed out of the total embedded systems. I don't think it is possible to know this. Too much of the information is close hold. I do know the percentage of failures varies by sector from a low of .5% to a high (I've seen) of 7%. I do know of many organisations (customers) who have exercised due diligence.

I regard the "refinery explosions, the mci failure, the power outage in chicago, the passport screwup in britain, the sewage dumpage in LA, the power outages" as evidence that people are making an effort to fix things - a positive.

The engineers fixing things don't go to the papers and say "I tested 40 things and two have to be fixed." This information doesn't end up in the media.

I think the problem here is the politicians and BJ artists in organisations are the ones who go to the media and say things like "we ran a test and everything is OK" or "We are Y2K ready", etc. Everyone who has dealt with this kind of character has an internal BS detector that goes off. So we conclude we are being lied to. We are, but that doesn't mean that many organisations are not doing diligent analytical, corrective and quality assurance. You just can't tell which ones aren't.

Everything is not going down. You just can't see through the B.S. outside of your own organisation and any you deal with directly, to see what is going down.

-- ng (cantprovideemail@none.com), August 24, 1999.

CygnusXI,

The Alliant energy rep and the USWEST rep confirmed last night that "TEST" are NOT conducted in a live environment between utility and telecommunication companies. TSHTF about 6pmEST on 21/31/99. No way around it, no end to end LIVE testing. You cannot TEST for ALL possibilities in a test lab(Marie will confirm this).

-- y2k dave (xsdaa111@hotmail.com), August 24, 1999.

Look, to suggest that there is only a 7% failure rate as a high is horseshi**. Entire systems have to be rolled over to ensure total compliance, and that ain't gonna happen folks. Why? Because when small parts of systems, be it electrical etc. are rolled over, they FAIL. Even pseudo tests on large, simulated systems, that apparently should work DON'T. So who in their right mind would risk the whole grid? The parts can't be remediated, so LOGICALLY, the WHOLE can't be remediated. This isn't hard. This is really quite simple.

Owl

-- owl (new@new.com), August 24, 1999.

Gosh, where oh where is MARIA when we need her? She knows ALL ABOUT testing. And stuff. She REALLY does!

Ohhhhh, Marrrrriiiiiiiiaaaaaaaaa!!!!!!

-- King of Spain (madrid@aol.com), August 24, 1999.

Sigh

"Look, to suggest that there is only a 7% failure rate as a high is horseshi**."

That was .5-7% varying by industry. Hundreds of them in a small factory. Tens of thousands on an military base (not counting weapons). They have to be dealt with individually. Sometimes one will stop a factory or topple a oil rig, but you can't work with this problem if you don't deal with them discretely, not a big conglomerate of a system.

"Entire systems have to be rolled over to ensure total compliance, and that ain't gonna happen folks. Why? Because when small parts of systems, be it electrical etc. are rolled over, they FAIL."

Not true in assessment, renovation, and testing. Lots of "stuff" can be rolled back. It may be aware of time as yymmddmmss, but it may need only to measure time, not have a need for accurate year. It is true that lots of systems operate interdependently, but lots are independent. Sometimes you can make interdependent ones independent. May degrade operations, but you can still work.

"Even pseudo tests on large, simulated systems, that apparently should work DON'T."

Some don't and some do. You just don't hear about those that do (except mindless blather from politicans).

"So who in their right mind would risk the whole grid?"

I assume you mean who would roll everything forward and test the power grid. That's what is going to happen in 129 days. So...

"The parts can't be remediated, so LOGICALLY, the WHOLE can't be remediated."

This is what you have to do if you are working on this problem. Decompose it into parts and work on the parts. That is why the first step in everyone's Y2K methodology in INVENTORY.

"This isn't hard. This is really quite simple."

Could've fooled me. This is the hardest and most complex technology problem I've worked on in my life - but "stuff" is getting done. Everything will not fail. I think Jim Lord's report is pretty close to accurate. Lots of places OK. Lots of places with one utility going down and having to be brought up. Some places in real trouble.

-- ng (cantprovideemail@none.com), August 24, 1999.

If you're trying to refute my statements, you're going to have to try harder than that. Where did you pull 7% out of, thin air? Based on what accurate reporting info? Do you actually believe the statistics? Come on, who will admit failure. The whole point about Y2K is that eventually, everything boils down to interdependence. Sure you can isolate a few insignificant programs, but the majority of mission critical systems run entire networks. And the REALITY of the situation is that when an isolated system fails upon remediation, and upon reremediation, there is no way an entire network will be rolled FORWARD. Besides, writing code isn't a science. Look how long it took to get Windows 95 up and running. There still debugging it. And Lord's report is a best guess, BASED ON NO REALTIME TESTING. Sorry, but ITS GOIN DOWN.

owl

-- owl (new@new.com), August 24, 1999.

I am trying to provide some reality to this discussion from expereience at actually doing this work, as opposed to conjecture.

Your questions:

"Where did you pull 7% out of, thin air?"

Actual number was .5% (a customer with buildings they were concerned about) to 7% (a embedded system subcontractor's of ours experience with a fortune 500 company whose product sits on my desk as I type). There may be an environment with a higher failure percentage, but I havn't validatable experience of it.

"Based on what accurate reporting info?"

Far as I know, there is no accurate reporting to the public. Where the real engineering is done, and problems are found, in my experience, they get tracked and fixed and NOT TALKED ABOUT till fixed. The bureaucrats, politicians, lawyers and PR guys handle public reporting in the government and SEC reporting in corporations. I havn't seen "ground truth" where there is real engineering get out yet. I think that is the reason there are so many aliases on this board.

"Do you actually believe the statistics?"

I took a an engineering course in statistics once. One of the texts was "How to Lie With Statistics". Good book. "I did not have sex with that woman. Miss Lewinski."

There is real work getting done and everything is not going down.

-- ng (cantprovideemail@none.com), August 24, 1999.

statistics are like bikinis

what they reveal is not anywhere near as interesting as what they conceal

-- andrea (mebsmebs@hotmail.com), August 25, 1999.

So we're blaming the British passport screwup on y2k?? Here in Britain we have a long and proud history of public-sector computer project screwups. The London Ambulance Service's new system failed abysmally all the way back in 1992. The Home Office's new system for handling political asylum applications is screwing up because they moved into a new office, put all the old paper files in storage somewhere and then decided to see if their new computers work. Andersen Consulting are trying as we speak to sort out the screwup they've made of the new National Insurance ('Social Security') system. The new air traffic control centre at Swanwick is years late and nowhere near working. These screwups are caused by managerial incompetence, which I feel confident will continue to be the leading cause of new system failure well into the next century.

Computers fail, chemical plants catch fire, planes crash and phones die. These things have happened for decades, so let's not assume that every failure we see around us now is down to y2k. Mind you, I'm sure *some* may be. But not our screwed up new British government systems!

-- Public Sector Cockup. (abcd@efgh.ijk), August 26, 1999.