A little fairy tale. Has a happy ending, and everything... Cheers, Bob Hettinga --- begin forwarded text Date: Wed, 17 Sep 1997 19:13:41 -0500 From: Bill GL Stafford <springco@arn.net> Organization: Spring Management Company MIME-Version: 1.0 To: Robert Hettinga <rah@shipwright.com> Subject: Preparing the Remnant for the far side of the crisis Robert; This is a rather long post. I wanted for you to see it personally. It is important and I wanted to clear it thru you. If I don't see it posted I will know you did not chose to post it. Best personal regards, Bill GL Stafford Began fowarded mail:

From: jmcnally@bigdog.fred.net Date sent: Thu, 11 Sep 1997 10:42:20 -0400 (EDT) Subject: REMNANT REVIEW of September 5, 1997

Gary North's REMNANT REVIEW emailbonus: Matt. 6:33-34 year2000@garynorth.com

Preparing the Remnant for the far side of the crisis

Vol. 24, No. 9 590 September 5, 1997

I am hereby lifting the copyright of this issue of Remnant Review. This one I want you to send to your friends, neighbors, boss, Congressman, and anyone else who might want advance information on the end, at long last, of the 16th Amendment: vetoed by Year 2000 noncompliant computers. Photocopy it, print it, whatever. Then visit my Web site for full documen- tation (under "Government"):

THE ULTIMATE TAX REFORM: JANUARY 1, 2000

What I am about to report will verify what I have been saying all year. If this doesn't constitute proof, I don't know what can persuade you. From this point on, anyone who tells you that the Millennium Bug is not a big deal, or who says, "We'll just have to wait and see about y2k, there's no need to hurry," simply doesn't know what he's talking about. Ignore him.

On August 21, I stumbled into the most amazing government document I have ever seen. I had read a brief news story about a company that had applied for a contract to work as a subcontractor for the IRS in a restructuring of its computer systems. The IRS admitted to Congress last January that its $4 billion, 11-year attempt to modernize its computer systems had failed. Here was a follow-up story. So, I went to the company's Web site to find more information. This led me on a merry chase across the Web.

Finally I landed on the IRS's page -- specifically, its page relating to its PRIME project. There were pages of blue links to documents, each one with a strange name or the name of a state. It was not clear to me at first what I had discovered. So, I started clicking links. I found nothing that I could understand, link after link: government bureaucratese. Then I hit pay dirt: the mother lode, my friends -- what we have been waiting for since 1913. Deliverance. Free at last, free at last! THE IRS'S MAINFRAME COMPUTERS -- 63 OF THEM, PLUS MICROCOMPUTERS -- ARE ON THE BRINK OF TOTAL COLLAPSE. Yee-hah!

This amazing admission appears in an innocuously titled document, "Request for Comments (RFC) for Modernization Prime Systems Integration Services Contractor" (May 15, 1997). The author is Arthur Gross, Associate Commissioner of the IRS and Chief Information Officer, i.e., the senior IRS computer honcho. It was Mr. Gross who went before Congress in January to admit defeat.

Mr. Gross now says that the IRS is no longer capable of operating its own computer systems. The IRS has over 7,500 people involved in just computer maintenance, with a budget a $1 billion a year (Appendix B. p. 2), yet they can no longer handle the load. And so, says Mr. Gross, some of them are going to get fired. You can imagine the continuing morale problem that this announcement will cause! The IS (information systems) division will be, as they say, DOWNSIZED. From now on, the IRS must achieve the following:

. . . shifting the focus of IS management to a business orientation: servicing customers with exponentially increasing technology needs, implementing massive new technology applications on schedule within budget while managing the downsizing of the IS organization (Appendix B. p. 2).

Do you think that people slaving away in their cubicles, trying to fix the Millennium Bug, will respond favorably to this notice? "Fix it, and then you're out!" Mr. Gross knows better. So, with this amazing document, he calls on private industry to come in and TAKE OVER THE ENTIRE IRS COMPUTER DIVISION. This is what Mr. Gross calls "a strategic partnership" (p. 1). The new partners will have to fix the Millennium Bug. The IRS will give them exactly eight months, start to finish: October 1, 1998 to the end of June, 1999.

The IRS's Digital Augean Stables

Perhaps you have had trouble on occasion getting information from the IRS about your account. After reading this document, I now know why. The information is held in what the IRS calls "Master Files" (p. 4). These files are held in the Martinsburg, West Virginia, computer. This computer receives data sent in by 10 regional centers that use a total of 60 separate mainframes. These mainframes do not talk to each other. Or, as Mr. Gross puts it, they are part of "an extraordinarily complex array of legacy and stand-alone modernized systems with respect both to connectivity and inoperability between the mainframe platforms and the plethora of distributed systems" (p. 4). This is bureaucratese, but I do understand the word, INOPERABILITY.

The tax data build up in the local mainframes for five business days. Then they are uploaded to West Virginia. This may take up to 10 actual days. Then the Martinsburg computer sends it all back to the regional computers in the Service Centers. Then the information is made available to the "Customer Service Representatives" (p. 5), i.e., local tax collectors. The elapsed time may take two weeks.

But . . . it turns out that the actual source payment documents are not sent to the Master Files. Neither is "specific payment or tax information." This information stays in what the IRS calls STOVEPIPED SYSTEMS, meaning stand-alone data bases "which, for the most part, are not integrated with either the Master Files or the corporate on-line system, IDRS" (p. 5). Separate tax assessments for the same person can appear in six separate systems, and these do not communicate with each other (p. 5). "Further, each system generates management reporting information which is not homogeneous, one with the other . . ." (p. 7). To help us visualize this mess, and much larger messes, the document includes charts. These charts are so complex that my printer was unable to print out the 116-page document -- probably not enough RAM. I had to get two other people involved to get one readable copy.

I have included one of these charts on the back page, just for fun. Go ahead. Take a quick look. No need to get out your magnifying glass just yet. Then comes the key admission: "These infrastructures are largely not century date compliant . . ." (p. 11). The phrase "century date compliant" is the government's phrase for Year 2000-compliant. In other words, THE IRS'S COMPUTERS ARE GOING TO CRASH. Now hear this:

In addition to three computing centers, (Memphis, Detroit and Martinsburg) the latter of which is a fully operational tax processing center, the IRS deploys a total of sixty mainframes in its ten regional service centers.

None of the mainframes are compliant, thereby necessitating immediate actions ranging from systems software upgrades to replacement (p. 9).

It gets worse:

A still greater and far reaching wave of work in the form of the Century Date Project is cascading over the diminishing workforce that is already insufficient to keep pace with the historical levels of workload. For the Internal Revenue Service, the Century Date Project is uniquely challenging, given the aged and non- century compliant date legacy applications and infrastructure as well as thousands of undocumented applications systems developed by business personnel in the IRS field operations which are resident on distributed infrastructures but not as yet inventoried (p. 13).

Notice especially two key words: "undocumented" and "inventoried." "Undocumented" means there is no code writer's manual. They either lost it or they never had it. "Inventoried" means they know where all of the code is installed. But it says: "not as yet inventoried." How much code? Lots.

The IS organization has inventoried and scheduled for analysis and conversion, as required, the approximately 62 million lines of computer code comprising the IRS core business systems. With respect to the business supported field applications and infrastructures, however, we do not know what we do not know. Until central field systems and infrastructures are completed, the IRS will be unable to analyze, plan, and schedule the field system conversion (p. 13).

I love this phrase: WE DO NOT KNOW WHAT WE DO NOT KNOW. This is surely not bureaucratese. Now, let's put all of this into a clearer perspective. The Social Security Administration discovered its y2k problem in 1989. In 1991, programmers began to work on correcting the agency's 30 million lines of code. By mid-1996, they had completed repairs on 6 million lines (CIO Magazine, Sept. 15, 1996.) Got that? It took five years for them to fix 6 million lines. But the IRS has 62 million lines THAT THEY KNOW ABOUT, but they don't know about the rest. It's out there, but there is no inventory of it.

Consider the fact that they have not completed their inventory. The 1996 "California White Paper," which is the y2k guide issued by the IS division of the California state government's y2k repair project, says that inventory constitutes 1% of the overall code repair project. Awareness is 1%. So, after you get finished with inventory, YOU HAVE 98% OF YOUR PROJECT AHEAD OF YOU. Meanwhile, the IRS has not yet completed its inventory.

The IRS has led the American welfare state into a trap. The Federal government, like the U.S. economy, will be restructured in the year 2000. Most Americans will be in bankruptcy by 2001, but they will be free.

Meanwhile, the news media are all a-dither about the Clinton- Congress accord on taxes, which will balance the budget in 2002. As George Gobel used to say, "Suuuuuure it will." Who is going to collect revenues in 2000?

Please Help Get Us Out of This Mess!

The next section of Mr. Gross's report I find truly unique. When was the last time you read something like this in an agency's report on its own capacity? (The next time will be the first.)

THE CHALLENGE: THE INFORMATION SYSTEMS (IS) ORGANIZATION LACKS SUFFICIENT TECHNICAL MANAGEMENT CAPACITY TO SIMULTANEOUSLY SUPPORT TODAY'S ENVIRONMENT, EFFECTUATE THE CENTURY DATE CONVERSION AND MANAGE MODERNIZATION (p. 13).

This states the IRS's problem clearly: its computer systems are just barely making it now, and the Year 2000 Problem will torpedo them.

Mr. Gross then announces the IRS's solution: quit. The IRS has now admitted that "tax administration is its core business" and it will now "shift responsibility for systems development and integration services to the private sector . . ." (p. 54). But first, it must find some well-heeled partners.

"The IRS has acknowledged that its expertise now and in the future is tax administration." This means that "the IS organization must be rebuilt to preserve the existing environment and partner with the private sector to Modernize the IRS" (p. 13). I love it when someone capitalizes "Modernize." Especially when it really means "officially bury."

Then the coup de grace: "Any reasonable strategy to move forward, therefore, would focus on managing the immediate crisis -- 'stay in business' while building capacity to prepare for future Modernization" (p. 14). Then comes part 2 of the report:

The Next Eighteen Months: Staying in Business and Preparing for Modernization

Mr. Gross knows that there is a deadline, and it isn't 2000. It's months earlier. He has selected June, 1999. Most organizations have elected December, 1998. This allows a year for testing. Mr. Gross is more realistic. He knows late 1998 is too early. The IRS can't do it. (I would say that late 2008 is too early. The IRS has tried to revamp its computer system before.)

. . . the IRS must undertake and complete major infrastructure initiatives no later than June 1999, to minimally ensure century date compliance for each of its existing mainframes and/or their successor platforms. At the same time, the IRS must complete the inventory of its field infrastructures as well as develop and exercise a century date compliance plan for the conversion replacement and/or elimination of those infrastructures. (p. 19).

Then comes an astounding sentence. This sentence is astounding because it begins with the word, IF. (Note: RFC stands for Request for Comment.) IF THE INFRASTRUCTURE ANALYSIS BECOMES AVAILABLE, UPDATES WILL BE PROVIDED TO POTENTIAL OFFERERS TO ASSIST IN DEVELOPING RESPONSES TO THE RFC.

If...? IF...? He is warning all those private firms that he is inviting in to clean up the mess that they may not be given the code analysis. But code analysis constitutes the crucial 15% of any Year 2000 repair job, according to the California White Paper. Then, and only then, can code revision begin.

Meanwhile, the IRS system's code is collapsing even without y2k. The programmers are not able to test all of the new code. Mr. Gross calls this "Product Assurance." This division, he says, has "sunk to staffing levels less than 30 percent of the minimum industry standard . . .. ." This makes it "one of the highest priorities within the IS organization, given that, today, major tax systems are not subjected to comprehensive testing prior to being migrated to production" (p. 15). In short, Congress passes new tax code legislation, and the IRS programmers implement these changes WITHOUT TESTING THE NEW CODE. Now comes y2k. As he says, "the Century Date Conversion will place an extraordinary additional burden on the Product Assurance Program." I don't want to bore you, but when I find the most amazing government document I've ever seen, I just can't stop. Neither could Mr. Gross:

Regrettably, the challenge is far more overarching: to modernize functioning but aged legacy systems which have been nearly irreparably overlaid by and interfaced with a tangle of stovepiped distributed applications systems and networked infrastructures (p. 55).

I'll summarize. The IRS has got bad code on 63 aging mainframe systems, plus micros. It has lost some of the code manuals. It does not know how much code it has. It must now move ("migrate") the data from these y2k noncompliant computers -- data stored in legacy programs that are not y2k compliant -- to new computers with new programs. These computers must interact with each other, unlike today's system. Bear in mind that some of this code -- I have seen estimates as high as 30% -- is written in Assembler language, which is not understood by most programmers today: perhaps 50,000 of them, worldwide (Cory Hamasaki's estimate). Then everything must be tested, side by side, old system vs. new system, on mainframe computers, before anyone can trust anything. (This assumes that extra mainframes are available, but they aren't.) Warning:

Beyond the magnitude of the applications system migration, the complexity and enormity of the date conversion that would be required necessitates careful planning and risk mitigation strategies (e.g., parallel processing). While the risks inherent in Phase III may be nearly incalculable given the age of the systems, the absence of critical documentation, dependency on Assembler Language Code (ALC) and the inevitable turnover of IRS workforce supporting these systems, it is essential to plan and execute the conversion of the Master Files and its related suite of applications (p. 30).

I'll say it's essential! The key question is: Is it possible? No.

Can you believe this sentence? "The risks inherent in Phase III may be nearly incalculable . . ." What does he mean, "may be"? They ARE.

Meanwhile, Congress keeps changing the Internal Revenue Code. This creates a programming nightmare: coding the new laws. So, how big is this project? Here is how Mr. Gross describes it: "Modernization is the single largest systems integration undertaking in world eclipsing in breadth and depth any previous efforts of either the public or private sector. Given the fluid nature of the Nation's Tax Laws, Modernization is likely to be the most dynamic, creating even greater complexity and, in turn, compounding the risks" (p. 54). Many, many risks.

Two questions arise: (1) Who is going to fix it? (2) At what price? The answer? He has no answer. All he knows is that this project is so huge that NORMAL COMPETITIVE BIDDING WILL NOT WORK. For this project, the IRS is not saying what its "partners" will be paid. It's open for negotiation.

You may be thinking: "Boondoggle." I'm thinking: "Legal liability in 2000 larger than any company's board of directors would rationally want to risk, unless they think Congress will pass a no-liability law in 2000." Here is Mr. Gross's description of the special arrangement. Pay close attention to the words "competitive process." He bold faces them; I do, too.

Our challenge, therefore, is to FORGE A BUSINESS PLAN AND PUBLIC/PRIVATE PARTNERSHIP in accordance with federal governmental procurement laws and regulations ABSENT THE TRADITIONAL LEVEL OF DETAILED REQUIREMENTS TYPICALLY ESTABLISHED AS THE BASIS OF THE COMPETITIVE PROCESS (p. 60).

He calls on businesses to create a "DETAILED SYSTEMS DEVELOPMENT PLAN" (p. 60). He goes on: "In general, the IRS seeks to create a business plan which: Shares risk with the private sector; Incents [incents???!!!] the private sector to either share or assume the 'front end' capital investment . . ." (p. 60). Read it again. Yes, it really says that. THE IRS WANTS THE PRIVATE SECTOR TO PUT UP MOST OR ALL OF THE MONEY TO FIX ITS ENTIRE SYSTEM.

This is why the minimum requirement for a company to make a bid is $200 million in working capital. It has to have experience in computers. It must be able to repair 5 million lines of code (p. 70).

How complex is this job? The complexity is mind-boggling: a seven- volume "Modernization Blueprint." To buy it on paper costs $465, or you can get a copy on a free CD-ROM. Needless to say, I got the CD-ROM.

So, you think, at least the IRS is getting on top of this problem. Suuuuure, it is. The contract award date is [let's hear a drum roll, please]: October 1, 1998 (p. 73). How realistic is this? You may remember Mr. Gross's deadline: June 1999. So, he expects these firms to be able to fix 62 million lines of noncompliant code, if they can find the missing code in the field offices, even though the IRS has lost the documentation for some of this code, in an eight-month window of productivity. Social Security isn't compliant after seven years of work on less than half the IRS's number of lines of code.

The IRS is facing a complete breakdown. Its staff can't fix the code. The IRS wants private firms to pay for the upgrade and manage the computer systems from now on. It does not know how much code it has. It does not have manuals for all of the old code. It does not even know how to pay the firms that get the contracts: either by "contractually greed upon fees" or "pursuant to measurable outcomes of the implemented systems" (p. 61). It has called for very large and experienced firms to submit comments by October 1, 1997.

In short, the IRS does not know what it is doing, let alone what it has to do. It only knows that it has to find a few suckers in private industry to bear the costs of implementing a new, improved IRS computer system and then assume responsibility for getting it Year 2000-compliant between October 1, 1998 and the end of June 1999. ("There's one born every minute.") Here are 12 companies that have expressed interest: nderson Consulting, Computer Sciences Corporation, EDS Government Services (EDS is not itself y2k compliant), GTE Government Systems, Hughes Information Technology Systems, IBM, Litton PRC, Lockheed Martin Corporation, Northrup Grumman Corporation, Ratheon E-Systems, Tracor Information Systems Company, and TRW. The list is posted at:

http://www.ustreas.gov/treasury/bureaus/irs/prime/interest.htm Conclusion

It's all over but the shouting. The IRS is going bye-bye. Accompanying it will be the political career of Mr. Gingrich and the historical reputation of Mr. Clinton. Bill Clinton will be remembered as the President on whose watch the Federal government shut down and stayed shut down. First Mate Newt will try to avoid going down with the ship of state, but he won't make it. And as for Al Gore . . . . Well, maybe he can get a job herding cattle on the Texas ranch of his ex- roommate at Harvard, Tommy Lee Jones. Think of it: not "Gore in 2000," but GORED IN 2000. Mr. Information Highway will hit a dead end.

On June 30, 1999, the IRS will know that its computers are still noncompliant. On the next day, July 1, fiscal year 2000 rolls over on the Federal government's computers and on every state government's computer that has not rolled over (and shut down) on a bi-annual basis on July 1, 1998. Almost every state: about half a dozen will roll over on October 1, 1999.

In 1999, chaos will hit the financial markets, all over the world -- assuming that this does not happen earlier, which I do not assume. The public will know the truth in 1999: THE DEFAULT ON U.S. GOVERNMENT DEBT IS AT HAND. The tax man won't be able to collect in 2000. The tax man will be blind. Consider how many banks and money market funds are filled with T-bills and T-bonds. Consider how the government will operate with the IRS completely shut down. Congress hasn't thought much about this. Neither has Bill Clinton.

END -- ÐÏࡱá Content-Type: text/x-vcard; charset=us-ascii; name="vcard.vcf" Content-Transfer-Encoding: 7bit Content-Description: Card for Bill GL Stafford Content-Disposition: attachment; filename="vcard.vcf" --- end forwarded text ----------------- Robert Hettinga (rah@shipwright.com), Philodox e$, 44 Farquhar Street, Boston, MA 02131 USA "... however it may deserve respect for its usefulness and antiquity, [predicting the end of the world] has not been found agreeable to experience." -- Edward Gibbon, 'Decline and Fall of the Roman Empire' The e$ Home Page: http://www.shipwright.com/