Gartner Group Final Published y2k Report

greenspun.com : LUSENET : TimeBomb 2000 (Y2000) : One Thread
I think everyone should read this report. Linkmeister, if you would please.

http://www.gartner.com

-- FLAME AWAY (BLehman202@aol.com), August 18, 1999

Answers

It is expired on other links and this gives me a blank page. Maybe they had nothing to say.

Best wishes,

-- Z1X4Y7 (Z1X4Y7@aol.com), August 18, 1999.

The report is here:

http: //www.gartnerweb.com/public/static/y2k/00081966.html

-- Hoffmeister (hoff_meister@my-deja.com), August 18, 1999.

I know the Gartner Group is smart. Buy how they got to their conclusion "there will be no global recession" is beyond me. It doesn't seem to come from their own evidence.

-- Rick Knight (rickk@flash.net), August 18, 1999.
I wish to emphasize three basic, self-admitted problems with this Gartner Group report:

(1) Y2K world status data, either Gartner Group's or any others'(Cap Gemini, Caper Jones, etc.) is practically useless, merely anecdotal as Gartner Group is willing to admit. Hard, asset-based, granular, independently validated Y2K world status data are non-existant. Gartner Group doesn't have them, neither does anybody else. Maybe the US State Dept. Y2K world status report adds something valuable by October. As this Gartner Group report points out, so far we all lack the type of data required to predict specific technical and economic consequences of Y2K worldwide.

Unfortunately, Hoffmeister believes he does (please see "Debate Round 1" - Hoffmeister/Heller Debate) and attempted to sustain the basic underpinnings of his arguments on worthless bits and pieces of information. A total falacy IMO.

(2) The Embedded Systems problem, as undefined as still remains, is nontheless frightening. Probable percent failures vary, but taking into account that worldwide the basic philosophy has been Fix On Failure, we should all expect serious consequences thereof.

(3) The intrinsic weakness and bias of self-reported data admitted in this Gartner Group report makes me wonder how much do we really know about Y2K even in the USA, supposedly world leader in Y2K remediation and testing.

Take care

-- George (jvilches@sminter.com.ar), August 19, 1999.

George, assumption (4) in my post related to the estimated percentage of Y2k problems addressed.

The original post assumed 66%. This was not intended to be a precise estimate. Indeed, the whole analysis was not meant to give precise estimates, but to demonstrate a relative rate of errors. As it stated quite clearly at the top.

Again, you and Steve questioned this number. Fine. Even assuming half of all systems with Y2k errors go completely untouched doesn't change the conclusions.

Yes, the model has as a basic underpinning that some percentage of Y2k problems are addressed. And no, there are no definitive sources to turn to for this estimate. But it is undeniable some percentage of the systems have been and are being replaced and fixed. Your problem seems to be the percentage used.

I used the same sources that others, including Ed Yourdon, has used. I used Capers Jones estimates; you don't like those, even though he obviously has at least considered international systems.

I used the history of software implementations. The same source that Ed Yourdon uses in his analysis. You apparently don't like that.

I agree, the Cap Gemini survey is inconclusive, without access to the actual definitions. But it certainly does not negate the estimate.

You now claim the GartnerGroup report demonstrates I was off. Fine, let's take a look.

The report is based on over 15,000 organizations, in 87 countries. This figure seems to be the most concise status estimate:

In Figure 4 and the following four figures, the lines represent the spectrum of progress replies received (excluding statistical anomalies).


Source: GartnerGroup

Figure 4. Progress by Company Size

Level 5 is done with all systems; level 4 is done with mission-critical systems.

Even this report doesn't negate my estimate. I'm not saying that 66% or 50% of all companies are completely done. The assumption is based on the percentage of potential errors fixed.

Thus far, all I've seen from you is much hand-waving that the estimate is wrong. If you have a better estimate, and can back it up, fine. Or if you have some backup that the number is grossly overstated, fine. Let's hear it. Thus far, I haven't seen anything.

-- Hoffmeister (hoff_meister@my-deja.com), August 19, 1999.

Hoff, before I get involved in your answer, let me add that another troublesome finding in this Gartner Group report is the 15% defects found in already remediated code, somewhat less in remediated and supposedly fully tested code, but still troublesome IMO. What do you think?

Now getting down to the Gartner Group report , it can only be considered to be what GG says it is: a high level assessment tool. Not a single bit more. I would add that this report can only be considered as a VERY high level assessment tool. The reason is that it lacks granularity and asset-based analysis, as Gartner Group openly admits. This also applies to Caper Jones, Cap Gemini, and any other Y2K world status estimate published thus far.

Example from the trenches (I've been there Hoff, same as you. The point is that its different trenches, as you'll see hereinafter)

(1) According to this GG report, Argentina's technology dependency is low. The reason is that the indicators that GG has chosen for this categorization is (a) the number of registered patents in a given country and (b) the R&D expenditure vis-a-vis GDP.

Now then, Argentina has very very funny Intellectual Property Rights policy and laws by which, legally, patents can be made use of and turned into commercial production without consent or agreement from the legitimate patent holder. This is Pirate Land gentlemen. In case of doubt check it out with the Dept. of State, the Dept. of Treasury, the Dept. of Commerce, and the White House. As a consequence, very few patents are registered in Argentina, of course. And another undesirable effect is that Argentina's r&d expenditures are very, very low vs. GDP. Ultra low I would dare to say. Argentina is not alone though, as many other countries indulge in similar practices, China, etc.)

Result: According to GG criteria, Argentina has very low dependency on technology, which is blatantly false: Argentina, by any reasonable standards, has medium level technology development, not any less. Actually, during the past 8 years the privatization of state companies and agencies has meant the implementation of brand new technologies. A full 60% of households in Argentina are wired by cable TV and fiber optics, and per capita cell phone consumption is the HIGHEST in the world. Many people I know have and use more than one cell phone.

So then GG goes on to conclude that as Argentina has very low technology dependency (false) it has little Y2K problems to solve (false) and is thus doing reasonable well in remediation (false again). The truth is that Argentina is far behind its required Y2K remediation and testing. And as I frequently talk and discuss these matters with Gartner Group people in Argentina, I have to assume that they would agree with me on this matter also. I cannot speak on their behalf, of course, but I may suppose that the self-reported bias is the answer to many of these "miracles", as the GG report itself suggests.

(2) Hoff, take Brazil, another country I know all too well. Brazil's recently privatized state phone company Telebras is far from completing remediation, let alone testing, only sketched along flimsy lab tests scheduled for December 1999 that will probably never take place and won't mean much as we all know already (think MCI). Still, the GG report shows Brazil reasonably advanced in Y2K. But I can't imagine any country, any economy, working half right with screwed up phones... or maybe dead phones. And phones is just an example.

(3) Going back to Argentina for a moment, I find it hard to understand how any country or economy can work without banks, do you Hoff? Well this is the case simply because the "Banco de la Nacisn Argentina" is still a federal government bank that handles 50% of transactions and holds 20% of deposits in Argentina, currency of which is pegged to the US dollar by a conversion board, no more and no less. So any bank run, for whatever reason, would throw Argentina into chaos. The big problemo is that the Banco de la Nacisn Argentina has 12 million lines of Assembler code to remediate, without enough budget or IT resources even for mission critical systems, let alone testing. There are 35 other state-owned provincial and municipal banks in similar condition in Argentina, o.kay? So it doesn't really matter what happens to the rest of Argentina's banking system. If the Banco de la Nacisn Argentina doesn't make it, and if most of the other state-owned banks don't make it, Argentina's been.

(4) Now Hoff (and others) you might think: just let Argentina and Brazil go down the drain. Well, International Economics 101 says it ain't that simple guys because as of 1994 Brazil and Argentina have established a Customs Union (similar to Europe's) and if their "Mercosur" trade block turns Merco'sour', then that pretty ice-cream cone you see South of the Equator every time you look at a world map would go down the drain AND with it the markets and the profits (60- 80%) for US multinationals, a 500 b-billion foreign debt unpaid, their resources unavailable, etc.

So, in summary, without this granularity and specific asset-based analysis all of these "thumb rule" estimates of code to be remediated and percentages of completion/remediation/testing/bugs/et al are simply worthless. That's why I think that your posting Hoff in "Debate: Round 1" has very weak underpinnings and should not be taken into account.

Take care.

-- George (jvilches@sminter.com.ar), August 19, 1999.

Hey Flint, where are you? I miss you guy.

-- George (jvilches@sminter.com.ar), August 19, 1999.
Hey, where is everyone? Poole, Mr.Decker... Marma maybe?

Guys, this is THE Gartner Group report you know, supposedly the whether bell of Y2K remediation. No comments anyone? Hoff? Flint? It's been a full day since my last post and surely it isn't that hard to come up with an answer, is it? Or does everyone agree with me this time ?

Take care

-- George (jvilches@sminter.com.ar), August 20, 1999.

George is correct. It is painfully necessary to analyze each country in detail, in depth and in specific (including embeddeds) and then to relate the implications of readiness-or-not with respect to their integration into a variety of supply chains (local, regional, global).

Most analyses are intensely weighted towards U.S. and a few partners, with some multinational corporate analysis (certainly vital) thrown in, all with complete U.S.-centric slant.

Analyzing function points is a useful part of the game but only one aspect and a rather minor one. Y2K is a game that involves politics and culture alongside of system failures (this will be especially true as we near rollover and go beyond) and the Gartners/Meta/Forresters are totally clueless there.

I'm hoping to get enuf time this weekend to read the "debate" and comment on the narrow technical stuff.

Meanwhile, George, keep at it and stay focused on what YOU know, which are the details down "south". An even more detailed analysis is, alas because of the implications, the best rebuttal, though indirect, to the polly side of the debate.

-- BigDog (BigDog@duffer.com), August 20, 1999.

I'd wager the CIA report is more accurate than the Garter Grope effort.

-- Andy (2000EOD@prodigy.net), August 20, 1999.
Sorry, George. Was a'travelin yesterday.

Agreed, any attempt at worldwide assessment must be at a very high-level. Again, this was an attempt at comparing relative error rates, not precise estimates.

You may want to re-read the GartnerGroup report. The criteria you based your answer on were (a) the number of registered patents in a given country and (b) the R&D expenditure vis-a-vis GDP..

But, according to the report, these are just:

Examples of additional indicators that enterprises can use to evaluate risk in various geographies and sectors include:

And in fact, they are not the indicators GartnerGroup used in the Report to determine the level of technological dependance. From the report:

For the purpose of this Strategic Analysis Report, we have selected the "World Competitiveness Yearbook 1999" by the International Institute for Management Development (http://www.imd.ch/wcy/wcy1999.html) as a single source of reference. The parameters we have considered are:

I'll look at the rest later. Just didn't want you to think I was ignoring you.

-- Hoffmeister (hoff_meister@my-deja.com), August 20, 1999.

Thanks for your input BigDog. No matter how hard I tried, I couldn't have said it any better myself. You have defined with precise but yet simple wording one of the deadliest sides of the Y2K animal.

As far as to keep focused down South, coun't on it. I know this side of the Y2K turf like the back of my hand.

BTW, middle-of-the-road supply chain problems in Latin America could still have very serious impact upon the US economy. What I mean is that serious problems could arise even without the banking system or the phone system in L.A. going haywire. Shortage of key mineral products and raw materials would do it. Revenue and profits for US multinationals could take a very heavy hit even without South of the border TEOTWAWKI.

Hi Andy !

Take care guys

-- George (jvilches@sminter.com.ar), August 20, 1999.

Hoff, thanks for your answer and please excuse my lack of patience.

This time though you are providing "smoking gun" evidence which proves my point even further.

For Brazil, Argentina, Venezuela, and Latin America at large (maybe with the partial exception of Chile, definetly the best student in the class) :

(a) "number of computers and number of telephones per 100 inhabitants (measuring information and communication technology penetration"

Answer: Very low, with the exception of Argentina's relatively recent telecomm development.

(b) "energy consumption per capita (measuring industrial intensity)"

Answer: low/medium, depending upon the country. Paraguay and Uruguay would be very low, Brazil and Argentina would be medium.

(c) "the trade-to-GDP ratio (measuring international dependencies)"

Answer: ULTRA low

So by these criteria, again, Gartner Group or anyone else would be mislead to believe that little Y2K remediation was needed. Simply not true. It's just that these indicators are way too unrefined. They lack the resolution needed to measure the subject matter at hand.

'Cause you see Hoff, in a nutshell, there are no bulky middle classes in Latin America. This is wild territory by US-centric standards. Broadly speaking, people don't get to have computers and phones nor many other things, with the funny exceptions of TV sets (wonder why?). Still, banks, insurance companies, power stations, government agencies, refineries, oil production, steel industry, etc., have a medium IT involvement which requires Y2K remediation which has not been done. Get the gist Hoff? Latin America still continues to have relatively underdeveloped economies WITH a medium IT dependency (particularly Y2K-wise).

After having travelled around the world more than a million miles I have the feeling that back home the US slant is so biased that most Americans have a highly distorted view of what the rest of the world is all about.

Take care

-- George (jvilches@sminter.com.ar), August 20, 1999.

George, what is your view, if you have one, on the status of seed supplies and ability to deliver from South and Central America to U.S. next year and/or 2001?

-- BigDog (BigDog@duffer.com), August 20, 1999.
Big Dog,

I don't really know anything specific about current seed trade data between Central and South America and the US. I could find out though.

Nevertheless, hybrid seed technologies and others are around and have been applied in some areas, for export purposes also (Pioneer et al). Still,I ignore domestic and export volumes of consumption.

Having said that, certain things are key under Y2K scenarios. Farming, seeding, etc., is definetly extensive throughout Latin America. It is NOT intensive. This means it depends upon natural rainfall, not irrigation, which is both good and bad, depending upon luck. Feed lot doesn't exist (only very little feed supplements for dairy cattle). Natural pastures is 99% of feed around here. Thanks to that and to other not-so-important factors, this area of the world is BSE free (Bovine Spongiform Encephalopathy or "mad cow disease"). BIG BIG deal in Europe and getting to be a big deal in the US in the near future BTW.

Storage capacity is very small, that's why Latin American farm exports (including seeds) are at the mercy of the market (whatever that means under Y2K conditions).

Farms themselves get power mostly from diesel gensets. Tons of manual labor still. Pretty self-sustainable up to 5-6 Y2K scenarios. If no diesel oil is available, I don't know...!

Take care

-- George (jvilches@sminter.com.ar), August 20, 1999.

I see, George.

You first state the Gartner report is wrong on their technological dependancy, and give as an example per capita cell phone consumption.

I point out that they communications is, in fact, one of the indicators that Gartner used in the report to measure technoligical dependancy, and somehow I've provided a "smoking gun".

Again, any form of encompassing analysis has to make some basic assumptions, assumptions that in individual instances may or may not be valid. I have no specific information on Argentina or Brazil, and have no way of either validating or invalidating your reports.

But it is no more valid to extrapolate to the whole from individual instances of supposed "failure", than it is to extrapolate from individual instances of "success". Which is why I did not base my original analysis on my personal experience.

This was an attempt to look at overall error rates, and compare them. It is not a replacement for individual assessment of individual situations.

-- Hoffmeister (hoff_meister@my-deja.com), August 20, 1999.

So Hoff, for lack of a better tool you are using a worthless tool !

It's like using a telescope trying to find the type of thing you would only find with a microscope.

And then, using the telescope, you go on and on shaving the unshavable shavings forever, you process them, you reach conclusions, forgetting they are worthless Hoff. Listen Hoff, I and everybody else gets a pretty good idea of how knowledgeable you must be in your field of expertise. It's nothing against you personally. We don't even know each other Hoff. As a matter of fact, you could be a computer or a committee for all I know (probably a committee which you lead, but that's a different subject). Now then, the methodology you follow could certainly be applicable to known IT situations, plus or minus whatever error, which you always openly admit, etc. That's all fine with me.

But Y2K is a completely different animal, a completely new paradigm. That's why you and many others have a hard time seeing it coming in its full magnitude and dimentions. Like I wouldn't want to challenge the validity of the estimated amount of code that has to be remediated and tested in the US, nor how critical it is here or there. I have ball park ideas of where the ball would end up landing, but the many parameters involved are way too complex and beyond my scope of understanding to get into deep analysis. Now, of course, if the Navy/Lord report is 10% true we are in deep doo-doo anyway. That I do understand. We don't need to be rocket scientists for that right?

But because of first hand knowledge, common sense and asset-based, case per case, critical path thinking and supply chain considerations, I can assure you the enormous risks involved in Y2K remediation and testing in Latin America, furious consequences included. I've given you a couple of examples above. It could present others. Latin America is smaller, simpler, more primitive and transparent in many ways, far more complex in others. Like Argentina doesn't have a thousand gas/oil pipes from fields to refineries and cities. It has three main ones. And if one of them happens to be way, way behind in SCADA remediation because I know first hand how delayed the bidding process still is (they haven't even started to fix or change anything yet-Aug/Sept 99), I can reach valid conclusions related to the serious risks involved, and so could you, without even caring about processing Cap Gemini data. To make things worse the number two gas/oil pipeline run out of budget three months ago because the provincial government involved doesn't care any more because polls show the incumbents will not be re-elected next October and by December 10 they would all be home, etc., etc., etc. Needless to say Hoff, no oil means no fuels, and no natural gas means no power, simply because 65% of power stations in Argentina run on natural gas. And BTW, all three pipes are pressurized by compressor stations that run on electric power. So if power outages take place, this means less natural gas or none at all, which means less electric power or no power at all. A perfect vicious circle that Satan himself could not have designed any better. You see Hoff, Gartner Group doesn't take that or many other things into account. Y2K is NOT only lines of code, IT equipment and embeddeds, it is ALSO key dates, CPM, node sustainability analysis, politics, local culture, economics etc. Did you know Hoff that Argentina's economy is 85% cash. Yes, CASH. Only 20% of the workforce has any relationship whatsover with a bank, not even a savings account. None, nada. Local culture (two 5000% hyperinflation periods in the past ten years) means a completely different ball game re bank runs. Completely different from probably everywhere else. Brazil is different. They have many other serious problems too, but national pride makes them keep their local cuurency much longer (not forever though) before cashing it in for US dollars, but not Argentines, they have developed ultra quick reflexes. Etc. etc., etc. Now taking these factors into account doesn't guarantee anything of course, but at least you are down to earth all of a sudden. Believing that Latin America has little Y2K remediation to do and therefore well ahead is false though, as I've said before.

The Gartner Groups and the Cap Geminis and the Caper Jones of this world can't help not having developed better indicators either. It is not their fault. But maybe their fault is not being humble enough to accept that their IT approach at the Y2K problem is so weak and unprecise it should not be even mentioned as evidence of anything, simply because it turns out misleading capable people like you Hoff. Let alone the enormous problem of biased self-reporting they openly admit.

So I believe the only possible approach to even address Y2K is to apply intellectual capacity, personal knowledge, and input from those you trust. Now I trust your brains Hoff, but I do not trust your methodology nor your underpinnings because it's like trying to find out which jockey will win the horse race by applying mathematical equations. Your chances are much better (however unprecise) if you knew the way the local horse races are rigged. In the case of Latin America, that's your ONLY chance. And I wouldn't be surprised if this also holds true for other parts of the world, including the US, depending upon the State (Arkansas?) or the industry sector.

Take care

-- George (jvilches@sminter.com.ar), August 20, 1999.

Hoff, I have three questions for you, if I may:

(1) Mr. Hoptiak has emphasized the importance of data-file and data- base remediation and testing, GIGO, cross-contamination, etc., supposedly "non-critical" and pretty much overlooked. (Please see his post at the "Debate: Round 1" thread, Hoff-Heller debate). What is your take on this Hoff ?

(2) What is your opinion on the Gartner Group's reported 15% defects in already remediated code?

(3) Having heard about the many tough problems that Russia has been facing for the past few years, what would be your educated guess on the status of Russia's oil and gas production and the Gazprom pipe which feeds into Germany 50% of the natural gas they consume (it's winter as you know)? Would you feel comfortable knowing that as of March 99 the Russians have guaranteed normal supply of oil and gas and that the Germans have taken their word for it without any IV&V?

Take care

-- George (jvilches@sminter.com.ar), August 20, 1999.

George:

1) As for GIGO, his points, as far as they go, are correct. Repairing data is by far harder and takes longer than actual code fixes.

But there is nothing to lead me to believe that Year 2000 errors create data file errors at any greater rate than the errors created during system implementations.

So comparing relative error rates is still valid. The resulting errors in data files are happening today, as well.

2) The analysis did take into account missed errors. I used Capers Jones' estimate of 5% missed; replace it with 15%, if you like. The analysis basically remains unchanged.

3) As for Russia and Germany, I'll have to look at this situation. Don't have any real opinion right now one way or another.

-- Hoffmeister (hoff_meister@my-deja.com), August 20, 1999.

Hoff, please read my lips and hold me 'accountable': By October 1, 1999, the US State Dept. will confirm my well-known forecast: Brazil, China, Russia, Italy and Indonesia will be the five most Y2K dangerous countries in the world. Many others would also run utmost serious Y2K risks, but these five will surely be amongst the top worse. The consequences for the world economy in general and the US economy in particular are enormous.

The problem I have with your low-level resolution methodology Hoff based upon broad, un-specific underpinnings, is that it doesn't reveal any of this at all. It leaves us clueless about concrete points/nodes of failures and says literally nothing about the impact upon the supply/value chains that affect us most.

After all, economics is the science involved in the creation and distribution of wealth, isn't it? Well guys, the economy is the jugular point of attack of the Y2K beast.

Take care

-- George (jvilches@sminter.com.ar), August 21, 1999.

Please check through the "Milne: Days numbered for the house of cards" thread posted today. It is quite relevant. Source: Reuters

Take care

-- George (jvilches@sminter.com.ar), August 21, 1999.

Moderation questions? read the FAQ