Noone seems to know about 9/9/99 as being a problem. The experts, even on the forum are divided. So lets use a little logic here:

1. Why isn't 9/9/99 understood? Because we can't really test for it without shutting down systems or taking them off line. So, using a little more logic,

2. If we can't predetermine if 9/9/99 is gonna be a problem cause we can't take systems off line, we sure can't determine if 01/01/00 is gonna be, now can we. This proves my point, that very little 'testing' has been going on, just a lot of fixing, with the hope that it will work. So the utilities are really lying by saying they're ready. They have fixed their systems, and they are ready for 2000. They just don't know if its gonna work, and that's what they've been saying. The real lies are coming from those saying to prepare for 3 days.

Owl

-- owl (new@new.com), September 01, 1999

Answers

made a few typos that renders the above message incoherent. Meant to say that the utilities 'aren't' lying by saying they're 'ready'. Also the reason things point to a catastrophe is because many systems I've read about that have been tested live have failed, or don't live up to expectations. But the logic remains the same, people shouldn't be deceived into thinking that actual repairs that 'work' are taking place. Because if that were possible, 9/9/99 wouldn't be mystery.

Owl

-- owl (new@new.com), September 01, 1999.

9/9/99 shouldnt be a problem, at least if you use good programming and a language which does handle a date string correctly.

It is possible that some programmers used 9/9/99 by parsing a date string turning it into numeric fields and then working with the month, date and year fields individually. It doesnt seem likely but its possible. Its going to just be a wait and see. At least on 9/10/99 we will know how many really bad programs we had still running until today.

I wish it were 9/10/99 so we can forget another one already.

-- parser (parser@chopthestringup.com), September 01, 1999.

September 9, 1999, would be treated as 090999. I don't see a problem there.

-- helen (sstaten@fullnet.net), September 01, 1999.

What about sabotage? Or, virus trip date?

Mike

=====================================================================

-- Michael Taylor (mtdesign3@aol.com), September 01, 1999.

I haven't seen a program with 9/9/99 in it since the late 70's, when I was writing a lot of PL/1 code. Most code based on PC's or Unix is going to be using a count-of-seconds style time stamp, since that's what the OS returns (and it's so much easier to work with!) And even in the ancient mainframe world, 9/9/99 was only used because it was visually striking. 12/31/99 would have been an equally valid "special" date.

-- You Know... (notme@nothere.com), September 01, 1999.

Owl, the experts aren't divided, just the views printed on this forum. Remember, pollies want to keep the dreaded nines alive so they can righteously shout "see, wrong again" when nothing happens.

Owl, are you in any way associated with computer work? Testing is going on like crazy. It's certainly not impossible to test stuff, it's just tedious and expensive to set up test environments and design the tests. We've been doing this for years - I've been working on Y2k remediation since 1994. It's not much different from other kinds of computer testing, which I've been doing since 1974.

9/9/99 is a non-event, a red herring, a myth. Read the threads: over and over you'll see people with 20, 30 years, saying "never saw it", saying "never heard of it". I have yet to see one person say "yes, I saw code that would break", the closest is a 30-ish year coder who said he/she GUESSED that it was POSSIBLE that someone might write a program that way, but that he/she never actually saw it or heard of it.

What does it take to set your mind at ease on the nines?

And your point is that very little testing is going on? Where do you work, what do you do? What gave rise to this opinion?

Look, I'm not saying 1/1/2000 is something to rest easy about. All the tests are in artificial environments, and we have to put a boundary around the test. We cannot build a complete factory just like a real one, fill it with exactly the same computers, and then roll the dates. (I heard of something close, but a nondisclosure agreement prevents my giving details.) So every test is designed to mimic real life, but it is NOT real life. We test all we can, but we'll KNOW on 1/1/2000.

One thing I agree on - the three-day lie will bite us. FEMA and Red Cross officials know this is a lie, they are individually prepared (many have told me) for a month minimum. No political leader, no emergency manager, no well-known business leader seems willing to step up and be honest on this one. Rotten shame.

-- bw (home@puget.sound), September 01, 1999.

Well well, now "You Know" makes two.

I think that's not enough to make me head for the bunker next week.

-- bw (home@puget.sound), September 01, 1999.

"Look, I'm not saying 1/1/2000 is something to rest easy about. All the tests are in artificial environments, and we have to put a boundary around the test. We cannot build a complete factory just like a real one, fill it with exactly the same computers, and then roll the dates."

Hi BW, this is my point. Testing in an artificial environment with boundaries is a pretty significant statement. I appreciate that programmers are racking their brains to try and fix this, but by 'fixing' this I mean FIXING it. Too much is at stake to hope that the artificial tests will suffice. And realistically speaking, and I would appreciate some really forthright honesty here, can we expect simulated tests to work in the real world. Somehow I really, really, really doubt it. I give it 0%. I'm trying to be completely pragmatic here. My background is irrelevant. I understand logic, and I definitely understand the difference between artificial and real environments. I'm a bit suprised that you're not stating the obvious - that whatever tests are being done will fail on rollover, because that's what initial tests do. Except, this time, we won't be able to fix it. A high stakes game of craps.

Owl

-- owl (new@new.com), September 01, 1999.

Hmm, I should have left out the remark about your background. I'm a little frustrated with the 9's question, but that remark is closer to personal attacks than I wanted to be. I hate that aspect of these discussions. Sorry.

Right, we won't *really* know what's fixed until 1/1/2000 (and stuff will break for months after that). It's the big unknown. The problem is made worse by the lies clearly coming out of major companies and government agencies. Maybe they fixed it and think so, maybe they fixed it but don't realize it, maybe they didn't but don't know it, maybe they didn't and know it and are lying about it.

So, to your points.

"Can we expect simulated tests to work in the real world"? Well, sometimes. Take Boeing. Pretty good at simulating stuff, so that when a new plane takes off for its first test it just doesn't crash any more. 40 years ago that was not the case. The computer geeks' ability to simulate the real computer's state is a bit spotty. Some are good, some are idiots. Top of my head, accounting for the range of programmers I've seen and the range of problems, the damage done by idiot managers, budget limits, etc, I'd say we are inevitably going to have major failures even in systems that are honestly thought to be ok.

"whatever tests are being done will fail on rollover, because that's what initial tests do". Often true, but "will" is too strong. Many tests will prove (on 1/1/2000 and later) to have been inadequate and the systems will break. I have (when on a coding high) written thousands of lines that have run correctly first time. It's rare, but it happens. Usually I muff something really elementary, and that's happening right now in Y2k work, I'm sure.

"this time, we won't be able to fix it". Absolutely do not know. This is the big question. How much will work, how much will break, and can we fix it before the mob breaks down the door? The "it" here is thousands of systems, hundreds of languages, some by wizards, some by idiots, being fixed on the fly by a conglomeration of wizards and idiots. "Able to fix" really can't be addressed without a time limit. Fix in a day? Fix in a year? For some systems, if you can't fix in a day you might was well lock the doors and go home. For others, maybe a month is fine.

We absolutely do not know what is gonna break, or how it breaks. As Halim Dunskey (sp?) said, we don't know if the system breaks like skin, and heals itself, or like a balloon that bursts.

"A high stakes game of craps." Yup, got that right.

-- bw (home@puget.sound), September 01, 1999.

BW,

Thanks for taking so much time to do your posts here and on the thread with Linda, who "feels like she is losing her mind. I felt like you had been eavesdropping on my musings last weekend; it was the same kind of thinking about the primal (probably a fundamental of human evolution) urge to belong to the family/tribe, for which we are hard-wired. Most of us are not accustomed to thinking and behaving so differently from our friends and families, and find it is a very stressful place to be for very long.

When I got to your posting here, and liked the sense and tone of it so much, also, I wanted to thank you for patiently and calmly, but respectfully, offering such a thoughtful contribution.

Puget Sound ... nice place to ride out whatever.. Good luck.

-- Kristi (KsaintA@aol.com), September 01, 1999.

I've never seen the 090999. The only thing close was Low Values (hex 000000000000) for headers or start of file and High Values (hex FFFFFFFFFFFF) for trailers or EOF.

-- MoVe Immediate (MVI@yepimhere.com), September 01, 1999.

Owl,

I don't know of a single programmer of any depth who expects 090999 to be a "problem". You don't sound like a geek (ette) by your post. Look back on North's forum for some background. Its called Y2K for a reason.

-- RDH (drherr@erols.com), September 01, 1999.

From: Y2K, ` la Carte by Dancr near Monterey, California

"You Know" said: I haven't seen a program with 9/9/99 in it since the late 70's

I was taught to use 8's, 9's, 98's and 99's for various different types of non-answers as late as 1984. That said, though, I don't expect next Thursday to have any noticable problems to the average person and certainly not to the news media.

-- Dancr (addy.available@my.webpage.neener.autospammers--regrets.greenspun), September 02, 1999.