Subject:	Gartner: Failures Rapidly Escalating, Barely Containing Them
Date:	1999/10/25
Author:	Paul Milne <fedinfo@halifax.com>
	Posting History

Y2K Is Months Away, But Bugs Turning Up Now Source: The Salt Lake Tribune According to the Gartner Group, only a small percentage of actual failures will occur with the change of the millennium. In fact, the pace of failures has picked up since Oct. 1, and will probably continue to rise into early next year. "At this point, most companies can handle the increased failures, but just barely," said Lou Marcoccio, a research director for the Stamford, Conn.-based firm. "The problem we have is that on Nov. 1, it will go higher and in December, dramatically higher," as computer programs come across more 2000 dates, he said. ======================== This is the reality of what is going on. The pace and severity of failures is escalating. Even the Gartner Group understands that there is a threshold of tolerance for these failures and that they are only just BARELY containing them now. What part of 'JUST BARELY' don't you understand. What part of 'DRAMATICALLY HIGHER' in December do you not understand? It is a statistical and historical FACT of IT departments that they do NOT reveal that they are behind until it is too late to do anything about it.. Preparationis only bears fruit if it is done BEFOREHAND. That is what the "PRE" part of 'PRE'pare means. It won't be much longer before the herd stirs and it will be too late. The admissions of failure are coming. The tangible objective failures are soon to be in full public view. -- Paul Milne "If you live within 5 miles of a 7-11, you're toast"

-- a (a@a.a), October 26, 1999

Answers

"a" - you're a bit late in the day. Check a few threads down to see one called

"GartnerGroups says the pace of failures has picked up since October 1st (Linkmeister, link@librarian.edu, 1999-10-25)"

(sorry, not hot-linked)

You can see Hoff and Big Dog circle around each other growling (but not snarling and biting yet)

-- Johnny Canuck (j_canuck@hotmail.com), October 26, 1999.

The original article is at this link:

http://cnniw.newsreal.com/cgi- bin/NewsService?osform_template=pages/cnniwStory&ID=cnniw&storypath=Ne ws/Story_1999_10_23.NRdb@2@18@3@143

-- Linkmeister (link@librarian.edu), October 26, 1999.

Fine with me "a", I'd prefer not to get involved in a thread with Mr. long-winded himself "Hoff, hoff, hoff, I'll blow your server down".

I think this statement should start to make a few pollys start to think for a change instead of fronting their false bravado...

"The problem we have is that on Nov. 1, it will go higher and in December, dramatically higher," as computer programs come across more 2000 dates, he said."

Pollyanna: "Duuuhh, oh yeah I forgot, it's da 2000 dates dat are da problem, not the 1999's!"

It is already starting to look like the rate of the disappearance of pollys from this forum is inversely proportional to the appearance of 2000 dates in computers. LOL!!

-- @ (@@@.@), October 26, 1999.

Thanks again for the post a.

Once again PM explains what should be the obvious in his clear to understand method.

I like never having to ask for a commentator to explain PM's point of view. If only he could get some regular air time, it could be educational and for some truly thought provoking.

Keep up the good work.

The meter is running and it wont be long until we will know the accuracy of predictions made by all sides in this

"The Mother of All Messes."

-- no talking please (breadlines@soupkitchen.gov), October 26, 1999.

You guys are too negative. Everything will be fine. Any problems will be easy to fix. The stock market will keep going up. The government said so and I believe everything I am told.

-- Dr. PollyDork (DrPollyDork@moron.com), October 26, 1999.

War is peace!

Celebrate Diversity!

-- Alan Stone (alanstone@eul.intermast.net), October 26, 1999.

Hey, how come Sissy Man hasn't piped up on this or the other Gartner Group thread with his standard "It's called the Year 2000 problem, not the Year 1999 problem!" Hypocrite!

-- Your basic troll (myemail@dress.com), October 26, 1999.

The @-person wrote:

I think this statement should start to make a few pollys start to think for a change instead of fronting their false bravado...

Hmmm.. 'false bravado'. Since bravado is 'a false showing of bravery', I guess you think that us pollys are truly brave. Better stop listening to old music and use a dictionary.

"The problem we have is that on Nov. 1, it will go higher and in December, dramatically higher," as computer programs come across more 2000 dates, he said."

Pollyanna: "Duuuhh, oh yeah I forgot, it's da 2000 dates dat are da problem, not the 1999's!"

Maybe you should look back and relaize that it was the doomers, not the pollys, that made all the noise about 1/1/99, 4/1/99, etc. ad nauseum. The reason was that, when those dates occurred, systems would start looking ahed far enough that some dates encountered would be in the year 2000. There was never a comment from pollys (or most doomers for that matter) that 1999 dates would cause problems. Only the more stupid doomers ever thought that 1999 dates (e.g. 9/9/99) would cause problems. Nice straw man you set up and knocked down, moron. Maybe you should try "to think for a change instead of fronting [your] false" intelligence.

-- Your basic troll (myemail@ddress.com), October 26, 1999.

Damn little slashes!

-- YBT (myemail@ddress.com), October 26, 1999.

Your basic troll,

You forgot the rest of my rubber stamp:

The number of mainframe, mid-range, PC, and embedded systems that do any look-ahead processing, is tiny when compared to the total number of systems that have a date problem.

In other words, the number of problems that we are experiencing now, is small when compared th the total number of problems that we will have.

They do call it Y2K for a very good reason.

What part of this don't you understand?

And as for Sissy Man, why don't you meet me out back after school. We'll see who the sissy is!

Tick... Tock... <:00=

-- Sysman (y2kboard@yahoo.com), October 26, 1999.

YBT,

My opinion on this is that a good portion of whatever problems there were that are related to fiscal year software and financial forecasting software have already taken place. But, many other types of prospective functions, let's say, making a doctor's appointment in December for January or ordering a part in late 1999 for arrival in early 2000 are still ahead of us. I could see this second type of problem increasing the closer we get to the end of the year.

Then, at or near the rollover, there's the potential for added problems with PC BIOS chip and PC operating systems, and embedded system/process control systems. That's when Y2K becomes more than just an accounting problem and could affect manufacturing, production or utilities.

-- Linkmeister (link@librarian.edu), October 26, 1999.

So what both of you are saying then is that The Gartner Group is wrong on this one? If not, please explain!

-- YBT (myemail@ddress.com), October 26, 1999.

YBT,

No, we're saying you're wrong and Gartner is right. Those 2000 dates are starting to roll in fast now little boy, so quit wasting time and start stocking up. It's gonna be one hell of a ride.

-- @ (@@@.@), October 26, 1999.

YBT,

What Link said, plus:

Yes, I do disagree with Gartner on this one.

Yes, I did disagree with the 1999 predictions, of Mr. Yourdon and others.

This doesn't mean that I think Gartner is crap. In general, they have done a pretty good job at the "best guess" game over the years. And we all know that I pretty much agree with Ed about Y2K.

As for Gartner, I do agree with their opinion on embedded systems, it will spike at 0 hour. And I do agree with their view that problems will "linger" well beyond 0 hour. However, I think that their prediction of early failures is too high, and their prediction of failures at 0 hour is too low.

But what do I know. I've only been messing with computers, from tiny to huge, for almost 32 years. My opinions are based on what I have seen, what I read about the problem, and my gut feeling. Who am I to challenge the great Gartner.

Tick... Tock... <:00=

-- Sysman (y2kboard@yahoo.com), October 26, 1999.

Oh well, no troll. I'm only bringing this to the top, because I was called on it, ie. "Sissy man". The punk chickened out, and didn't show up after school I see...

Perhaps one of his friends will take his place, and discuss the Gartner curve. I'm sure someone here can tell us all about it.

Oh, I see Hoffmeister is online this evening. What about it Hoff, care to discuss early failures, and how much "code" does forecasting?

Anyway, I be here...

Tick... Tock... <:00=

-- Sysman (y2kboard@yahoo.com), October 26, 1999.

Like I said on the other thread, I think Gartner's problem is not accounting for failures and errors from replacements and remediation. This statement strikes me as CYA, as in no-impact being felt to date.

I don't think forecasting problems are huge. Some production planning, automated ordering, etc. The financial reporting stuff may be affected, but nothing really external to a company.

My guess is Gartner's reasoning on Oct 1, Nov 1, etc., has more to do with billing and payment processing. You start hitting 90 and 60 day terms on payments, which start calculating due dates into 2000.

-- Hoffmeister (hoff_meister@my-deja.com), October 26, 1999.

Good evening Hoff,

"I think Gartner's problem is not accounting for failures and errors from replacements and remediation"

So, what are they accounting for? What other category does Y2K work come under?

"I don't think forecasting problems are huge"

Neither do I. I think it's tiny. What can we agree on? 5%, 10%, 25% of all date processing does any type of forcasting?

Yet, we do have early failures. "Companies Barely Containing Them" is an opinion, but they are dealing with them. What if 50% of all Y2K errors have already occured? Can we deal with it? Sure, no problem, been there, done that. But what if it's 25%? Can we deal with 4 times as many problems? What if it's 10%, and 10 times as many? What if it's only 5%, Can we deal with that? How many "percentage points" does it take, before this becomes a real problem???

My opinion of how CURRENT-DATE relates to Y2K is, I believe, well known here. It is, IMHO, the heart of the problem. Not future dates, but today's date, and yesterday's date. That's the current, real-time stuff. That's the stuff that most programs process.

It ain't Y2K yet, damn it...

Tick... Tock... <:00=

And a few PSes:

1) In our earlier discussion of the FAA, I implied that the FAA was hiding problems from GAO. That doesn't sound right. What I mean is that I find it very interesting that GAO found almost 1000 changes "in progress" that were not accounted for. And I find it hard to believe that this is normal maint work. They are, after all, doing a Y2K review, not a maint review. I just don't think the FAA is 100%, and I think they know it. Just a gut feeling.

2) I'm glad you're here tonight Hoff. I'm in a fightin' mood! I just learned in another thread tonight, that Robert X. Cringely isn't really Robert X. Cringely. Broke my heart... <:)=

-- Sysman (y2kboard@yahoo.com), October 26, 1999.

Gartner is accounting for problems with date processing. See http://www.gartnerweb.com/public/static/y2k/00081966.html#0016. They aren't accounting for non-date problems that occur due to system replacements and remediation.

Again, forecasting problems won't be much externally. But billing and payment processing problems will have an external effect.

-- Hoffmeister (hoff_meister@my-deja.com), October 26, 1999.

Again Hoff,

Forget Gartner. A problem is a problem. The question that I have is how much "code" deals with future dates? My opinion is that it's a small part, and the number of problems that we do see today, is also small. My opinion is that programs that do deal with forcasting would be fixed first, because they are needed first. Yet, we do have problems, now, related to Y2K. We still haven't reached CURRENT-DATE. We have virtually no year 2000 history in MOST of the systems, those that don't do anything with the future.

Come on Hoff, is this, or is it not the tip of the iceberg? How many "percentage points" of the iceberg are above water? <:)=

-- Sysman (y2kboard@yahoo.com), October 26, 1999.

Are you still out there, this early AM, Hoff? Or did you leave us for csy2k...

I gotta crash. I'll re-top this later. Maybe YBT will show up for class...

Since I like to dream warm things, how about we change this to "how much of the hot-air bag is full," instead of how much ice... ... ... <:)=

-- Sysman (y2kboard@yahoo.com), October 27, 1999.

PS - Maybe this would be a good time to talk about "testing" OK, good night, zzz, zzz, zzz...

-- Sysman (y2kboard@yahoo.com), October 27, 1999.

Sysman

I don't necessarily agree that a "problem is a problem". A problem in the ability to bill customers, or pay vendors, is more severe than an accounting forecasting problem.

I don't have any information that would contradict Gartner on their percentages, as they relate to strictly date problems. Like I mentioned above, billing and payment processing is starting to hit instructions that compare CURRENT-DATE to a year 2000 date, or is attempting to calculate a date that extends into the year 2000.

As for comparing rates of errors, that was the whole point of the analysis I attempted on replacement and remediation errors.

-- Hoffmeister (hoff_meister@my-deja.com), October 27, 1999.

Hoffmeister,

I do attempt to follow "pollies" opinions with an open mind, really I do. And whether I agree or not should 1. not matter to you at all, and 2.it is up to me, limited as I am understanding these things, to discern what you and others are saying.

However, you and other optimists (from other sources of information also) are really just staying one step ahead of all of these arguments, and loosing ground daily. That is not good. Heck, while I may more "agree" with Mr. Milne's "opinions", as example, do you really think that I am not "rooting" for "Hoffmeister(s)" to be right?

Now I do know a very little about accounting, though not accounting software (I used a pencil, manual typewriter and actually can use a memeograph machine!), but your last post: "A problem in the ability to bill customers, or pay vendors, is more severe than an accounting forecasting problem." is, well...

Is there an accountant in the house?!

Hoffmeister, you, me and the guy next door, we are all (most?) about to experience a very drastic change in our life (lifestyle?). Accept it for goodness sake, look beyond what you know, like I have been doing for almost a year. I don't know computers, but thank goodness I have always had the ability to connect the dots of a variety of aspects of our society (unfortunately that is not a marketable skill in the boondocks :-)

Thank you for being a person working on the problem in the "real world", from someone hoping alot of people are successful. But every day, reading here and other places, so many folks, need to rethink what they are predicting in words and how those words are going to actually pan out in the real world. "Connect the dots!" At least if a doomer predicts TEOWAWKI vividly, he/she can see the outcome correctly, i.e. translate prediction into real events. Some of these optimistic arguments only validate BITR on paper.

-- Lilly (homesteader145@yahoo.com), October 27, 1999.

Lilly

As for the accounting question, I'm basically agreeing with Sysman. Errors to this point would probably not be externally noticeable.

Problems in billing and payments will be more externally noticeable.

As for conecting the dots, being able to paint a picture of TEOTWAWKI doesn't make it any more likely to happen.

Forget the details of the error rate comparison for a second. Just look at the articles that Homer Beanfag has been posting lately. Virtually every one of them nails my point home further, though I doubt that is the intent.

We've been through a massive software upheaval, that generated enormous numbers of errors. No, they aren't all solved in "oh, 2 or 3 hours", as these articles describe. But these articles and errors are just the tip of the iceberg, compared to the number of errors and failures that occurred and were fixed.

And through all of this, we've survived quite nicely. Even prospered, to a certain extent.

That is reality, Lilly. That software errors have been dealt with; even at rates equal to or greater than Y2k.

-- Hoffmeister (hoff_meister@my-deja.com), October 27, 1999.

Italics off?

-- Johnny Canuck (j_canuck@hotmail.com), October 27, 1999.

I understand now, remediation vs. date problems. Another bit of learning for trying to figure out what the heck you guys mean sometimes! :-)

-- Lilly (homesteader145@yahoo.com), October 27, 1999.

Char-lie Brown do da dute do dute That Char-lie Brown do da dute dua dute He's a clown ... That Char-lie Brown

-- a (a@a.a), October 27, 1999.

OK Hoff, I'm back. Better late than never...

"As for the accounting question, I'm basically agreeing with Sysman."

Thank God. But I do have one point. Can you show me a "billing" system, that does a payment due date more than 30 days out? Better yet, show me one that does more than 25 days, allowing for processing time before the next close. So OK, then I've got a problem with this:

"Like I mentioned above, billing and payment processing is starting to hit instructions that compare CURRENT-DATE to a year 2000 date"

Sorry Hoff, this type of error isn't going to hit us for another 35 days, when the November closings start. Fact or fiction?

"I don't necessarily agree that a "problem is a problem""

Listen Hoff, I have studied your opinion. I know all about how much fix/replace work is going on, and all of the problems that come with it. Sorry, but I stand by my statement. It doesn't matter if it's a "date" problem or not. Here's the real question. Why is this new, or fixed, software here to begin with? Isn't it here to fix the Y2K problem? Aren't almost ALL "new" problems, with this "new" "almost does the job" software, really Y2K problems? I think we almost agree again here Hoff. Fact or fiction? (but I gotta tell ya again, you're just adding fuel to the fire)

This is amazing, we're almost in total agreement! Well, maybe not...

"I don't have any information that would contradict Gartner on their percentages"

Funny, neither do I. I remember looking at the Gartner chart (ya got me on this one, it isn't in my Y2K folder, like GAO is, so can ya post me a link Hoff, and save me the trip, pretty please...). Like I said, I agree with much of it. I do have a problem with, what was it, 7 or 9% failures, at 0 hour. It just seems to me that I've seen CURRENT-DATE in more that 7 or 9% of the places that I've seen do date calculations. And my other problem is their view of early failures. I think it's too high. In fact, I think that a few "percentage points" in the early failures category, belong in the 0 hour category. But hey, what's a few percentage points between friends. What do I know... Just a gut feeling, I guess. Neither fact nor fiction.

But I'm still not done, Hoff. You can tell me all about the work, and testing, and IVV that has been done. You can show me all of the happy-faced spin that you can find. The fact is, Hoff, that we are seeing problems now. We are seeing problems with the TINY PERCENTAGE of programs that do any kind of "forcasting" date math. Even after testing, even knowing that these programs would be needed first. We are seeing the tip of the iceberg. Didn't you just say that, Hoff?

Oh well, what can I say. We don't have much Y2K history yet. It ain't Y2K yet. Fact or fiction...

Tick... Tock... <:00=

-- Sysman (y2kboard@yahoo.com), October 28, 1999.

Oh well, I guess Hoff is over at csy2k. Yea, I know, I'm late. Damned life...

I guess I'll top it again later. Unless another polly would like to take over. I'll be here, maybe another half hour. 2 AM is my quittin' time... <:)=

-- Sysman (y2kboard@yahoo.com), October 28, 1999.

Hoff said he'd be back just as soon as he locates the Great Pumpkin.

-- Will continue (farming@home.com), October 28, 1999.

Sysman

"As for the accounting question, I'm basically agreeing with Sysman."

Thank God. But I do have one point. Can you show me a "billing" system, that does a payment due date more than 30 days out? Better yet, show me one that does more than 25 days, allowing for processing time before the next close. So OK, then I've got a problem with this:

"Like I mentioned above, billing and payment processing is starting to hit instructions that compare CURRENT-DATE to a year 2000 date"

Sorry Hoff, this type of error isn't going to hit us for another 35 days, when the November closings start. Fact or fiction?

Every system I've worked with (not just SAP) calculates and stores the "Net Due Date" when either a Customer or Vendor Invoice is entered. It saves processing, as opposed to re-calculating the due date based on terms on every run.

Which means that the data entry program has been forced to calculate a date into the year 2000, and store that date. Which also means the payment and billing runs have had to deal with year 2000 dates in selecting invoices.

Whether or not they are supposed to select only 25 days in advance isn't the point. The Y2k problems will occur IF they cannot correctly handle either the calculation of the Due Dates or Storing (Data Entry), or the comparison of dates (Processing).

"I don't necessarily agree that a "problem is a problem""

Listen Hoff, I have studied your opinion. I know all about how much fix/replace work is going on, and all of the problems that come with it. Sorry, but I stand by my statement. It doesn't matter if it's a "date" problem or not. Here's the real question. Why is this new, or fixed, software here to begin with? Isn't it here to fix the Y2K problem? Aren't almost ALL "new" problems, with this "new" "almost does the job" software, really Y2K problems? I think we almost agree again here Hoff. Fact or fiction? (but I gotta tell ya again, you're just adding fuel to the fire)

This is amazing, we're almost in total agreement! Well, maybe not...

In my opinion, it doesn't really matter whether or not replacement or remediation problems are regarded as "Y2k Problems". Except for the fact that Gartner doesn't include these in their percentage breakdowns.

Obviously, all "problems" are not equal. My only point here was that accounting/forecasting problems don't have the potential to externally effect business, as do billing and payment problems.

"Fuel to the Fire"? Maybe, if you mean that more errors have been generated than most thought. But the point is they've been dealt with. Even those that don't take, "oh, 2 or 3 hours" to fix.

But I'm still not done, Hoff. You can tell me all about the work, and testing, and IVV that has been done. You can show me all of the happy-faced spin that you can find. The fact is, Hoff, that we are seeing problems now. We are seeing problems with the TINY PERCENTAGE of programs that do any kind of "forcasting" date math. Even after testing, even knowing that these programs would be needed first. We are seeing the tip of the iceberg. Didn't you just say that, Hoff?

No, Sysman. I don't recall seeing any accounts of problems with forecasting, although I'm sure they occurred.

The "Tip of the Iceberg" remark was in regard to publicized problems with replacements. It isn't the "problems with the TINY PERCENTAGE of programs that do any kind of "forcasting" date math". It is the massive number of problems with the LARGE PERCENTAGE of systems that were replaced or modified, that have already occurred, and been dealt with.

-- Hoffmeister (hoff_meister@my-deja.com), October 28, 1999.

OK Hoff, I hear ya.

More to come. Coffee, shower, drive to work, check things out, more coffee... I'll be back. <:)=

-- Sysman (y2kboard@yahoo.com), October 29, 1999.

Good morning Hoff,

Sorry, late again. Got to go to our backup site in NYC again on Monday (bummer), and had mucho to do today.

Anyway, the good news. I agree that fix/replace is causing a large number of errors (not really good news, but I agree). I'll even go as far as saying I agree that "the Y2K problem" itself is less that 50% of the "total" problem, from your point of view.

But the reason for most, not all but most, f/r these days is a RESULT of Y2K. My mainframe would be fine for many more years, if not for Y2K. I'm far from alone. The fact is that we do have a fire. Y2K and f/r are both fuel. They will both cause errors. We all know much work has been done. If it has been enough, is the question.

Here's my problem Hoff. It's still early, but we are seeing early failures. From fixed, in production systems. Sure, we are dealing with them, and here's why: we have almost no Y2K data in database history. We haven't seen many problems yet, because most programs haven't processed Y2K data yet, in real world, production environments. Please, let's give it up with this one example from one system. We know how many examples are out there.

So yes Hoff, most code is in production, now. But most of that code cares about CURRENT-DATE, and history, like yesterday's date, or last weeks, or last months, not tomorrow's or next week or next month. By most, I mean a large majority. Fact of Fiction, Hoff? <:)=

-- Sysman (y2kboard@yahoo.com), October 30, 1999.

Sorry Hoff, another PS:

"Every system I've worked with (not just SAP) calculates and stores the "Net Due Date" when either a Customer or Vendor Invoice is entered. It saves processing, as opposed to re-calculating the due date based on terms on every run."

Please explain this "saves processing" to me. I've been doing SA work since the mid-70s, and every billing system that I've ever worked on, or designed, has something called a month-end, or closing. At this time, accumulated monthly credits and debits are "applied" to the statement, interest may be computed, the "next due date" is computed, the bill is sent out, etc.

Now, if you're talking about something like a credit card account, where you have a minimum balance, and a "projected monthly balance if you make minimum payments" or a "budget payment plan" or something like that, then hey, great, we've got another example of where "forcasting" works! Bring on the marching band, we have a successful Y2K fix!

But this has nothing to do with what John Doe sends in his payment envelope, and how it's processed. The budget payment plan has no impact on the business, except for a few bucks interest from the late payers, if they even pay it...

Business runs month-end to month-end, and we haven't had a Y2K month-end yet. We're gettin' close, but not yet. Yea, a few things here and there, but not yet!!!!!

Once again, let's try to forget individual examples. Can we come to an agreement? Doesn't MOST date processing deal with current or recent history, compared to future "forcast" dates? Can we at least agree on this?

And doesn't "mission critical" mean what shows up on the bottom line THIS MONTH, not what was predicted six months ago???

It still ain't Y2K...

Tick... Tock... <:00=

-- Sysman (y2kboard@yahoo.com), October 30, 1999.

Call the system replacements and modifications "Y2k-related" or "because of Y2k", or any other term you like.

Yes, errors due to "date" processing itself will rise.

Errors due to "replacements" and "modifications" have already peaked.

The point was to compare the two, one of which has already occurred, causing little or no impact to the vast majority of people. Not even a recession.

As for the closing stuff above, yeah, I'm fairly familiar with fiscal period closings. The point is most billing and payment processing is based on the "terms", some of which get fairly complicated. I used a simple example of "Net" due date; the date by which, if paid, no interest or penalties is charged. Other terms will include discounts, etc., but all are based on a given date. This date is calculated and stored with the invoice, both customer and vendor. Billing and payment runs compare either current date or perhaps fiscal period ending date to this date, to determine of payment should be made, or penalties charged, etc.

And again, I think the reason that Gartner expects the rise beginning in Oct is that terms such as "Net 90 Days" start becoming prevalent. So entry of the invoice involves calculation and storage of a date 90 days in the future, into the Year 2000. Billing and Payment runs start processing these dates, to determine invoices to be paid, or interest to be charged.

-- Hoffmeister (hoff_meister@my-deja.com), October 30, 1999.