This is a reply to an article written by Harry Browne about Y2K in World Net Daily.

Dear Sir or Madam:

  I have read the article by Harry Browne in the January 11th edition of
your publication and felt compelled to respond because I am uniquely
qualified to do so.
  You see, he classifies those who consider Y2K to be a serious problem
into two categories, both of which he dismisses as follows:

               "The Y2K problem has been exaggerated
               by people who don't understand
               computers, and by computer experts who
               don't understand how the free market
               works."

However, neither of those categories applies to me. While many people
can claim to be computer experts (with varying degrees of justification),  
very few people have been hired by Mr. Browne to modify programs he has 
written. I am in the latter category, having written assembly language 
routines to improve the performance of his home-grown word processor 
for the HP 9845. Given this, as well as my resume, I
don't think he can dismiss my expertise in the computer field.
  As for my understanding of how the free market works: I was introduced
to free market economics by Mr. Browne's books and have read quite a bit
about it since then, including many of the works that Mr. Browne himself
learned from. I would be willing to pit my knowledge of economics
against anyone's.
  And as long as we're talking about economics, I would like to point
out something that Mr. Browne seems to have overlooked in his analysis:
the companies that are the most critical to the survival of our
civilization are precisely those that are the farthest from the free
market. I'm referring to the utility companies, specifically water,
power, and telecommunications, as well as the banking system. All of
these are extremely heavily regulated by the government, and therefore
are generally very slow to adapt to changing circumstances. This is
undoubtedly one of the reasons that they have started working on the
year 2000 problem very late, and in some cases have barely begun.
  But there is another flaw in his argument that Y2K will not be a
serious problem because of the wonders of the free market. As big a fan
of the free market as I am, it still consists of people, millions of
people, who have all the frailties that humanity suffers from. One of
these frailties is an unwillingness to face unpleasant realities, and
another one is is a tendency to procrastinate. The former of these
frailties is the reason that it is so difficult to convince people to
prepare seriously for any disaster, whether it is Y2K or a hurricane
that is more than 12 hours away.
  The almost universal tendency to procrastination is the key to the
next fallacious statement in Mr. Browne's article:

	       "Many large companies do need to upgrade
               old computer systems and databases
               (although your personal computer probably
               will have no problems). Upgrading a
               computer system is a formidable task. But
               so is moving into a new factory, changing a
               product line, or dealing with new
               regulations. Companies deal with such
               problems as they arise, and one way or
               another they usually solve them."

  What is wrong with this statement? Just one missing word:
"eventually". It is well-documented that most large software projects
are delivered very late, and many are never delivered at all, being
cancelled before they are finished. (See Ed Yourdon's
article on this topic.
  In this case, of course, being late is disastrous, while cancellation
is suicidal, but those are the most likely outcomes of such a large
project. What is the conclusion? That most companies aren't going to
finish on time.
  The next fallacy in his article is the notion that because Y2K is a
widespread problem, that makes it easier to fix:

               "The Y2K problem seemed uniquely
               dangerous because millions of companies
               have to deal with it at the same time.
               Hundreds of thousands of COBOL
               programmers would have to be found -- to
               examine old computer programs, change
               every date reference, and test the
               corrections. But, in truth, a widespread
               problem is easier to handle, because it
               offers bigger profits to people who can
               devise solutions. 

               So now there are products like Revolve,
               Restore 2000, Milligration, and dozens
               more -- computer programs that go through
               old programs, fix the date problems, and
               test the results. These automated solutions
               eliminate the need for thousands of
               programmers."

This seems to be an excellent solution, but it doesn't explain why some
large companies like Citibank are spending hundreds of millions of
dollars on fixing their programs. The answer lies in the analysis of his
analogy between fixing Y2K and creating Web pages:


               "The Internet flourished in a similar,
               unpredictable way. If in 1994 someone
               had said there would be millions of World
               Wide Web sites in 1999, you might have
               assumed he didn't understand computers.
               Websites are written in a complicated
               computer language called HTML. Where
               are the hundreds of thousands of HTML
               writers necessary to build millions of
               sites? 

               But software companies came forward
               with computer programs that enable
               people to build websites without
               understanding HTML. Other programs
               help specialists to produce the more
               sophisticated, animated, interactive sites.
               The result is that we do have millions of
               websites after all."

  This comparison indicates a complete lack of understanding of what
"the Y2K problem" actually is. It is not one uniform problem that can be
fixed by any particular single solution, as is the task of creating a
program (or even a few programs) that can translate a user's description
of a Web page into HTML code. Instead, it is a collection of hundreds of
millions of individual problems, each of which is different from the
others. The difficulty is that there are almost an infinite number of
ways to perform the same task in a program; the HTML creating programs
that Mr. Browne refers to can do it in any way they wish and still get a
valid result, but fixing Y2K bugs requires that a person (or a program;
in some cases) figure out exactly what the program being fixed is doing
and how it is doing it. This can be quite difficult even for a small
program that is poorly written, much less for gigantic COBOL programs
that have been modified many times since they were written, by many
programmers.
  In fact, Y2K is not even entirely a COBOL problem as Mr. Browne
implies: Y2K problems exist in hundreds of computer languages in
hundreds of millions of programs around the world. The Department of
Defense alone uses more than 500 computer languages, programs written in
any of which can have Y2K problems. That is probably one of the reasons
why they're so far behind on fixing the problem, as indicated by the
extremely low "grades" that they have received from Stephen Horn,
Chairman of the House Subcommittee on Government Management,
Information, and Technology.
  But the real "killer application" of the Y2K problem is the one that
Mr. Browne doesn't even mention: embedded systems. These are the
computers that control much of our nation's infrastructure. They're
hidden inside oil pipeline machinery, refineries, power distribution
networks, generators, oil tankers, telephone systems, water treatment
plants, sewage treatment plants, and many other pieces of equipment on
which our life depends. The good news is that most of these embedded
systems will work properly when the clock rolls over to January 1st,
2000. 
  The bad news is that "most" is not enough. It has been estimated
that several percent of the systems will fail at rollover or shortly
thereafter. (See the following PC Week article). 
  That doesn't sound like very many failures, but it is more than
enough to bring industrial civilization to a halt. The difficulty is
that even one malfunctioning system in an industrial plant can cause a
disaster. On a number of occasions, managers of industrial plants have
set the clocks forward to see what would happen when rollover
occurred. In most of these cases, the plants ceased to function, and
in one case there was physical damage to the plant that took weeks to
fix. See Raleigh Martin's WWW site for details.
  Am I being overly alarmist? Not according to the National Guard's Web
site, which says:
"Large segments of our nation's electric power grid could fail, causing
massive blackouts. Water distribution systems could fail.  Distribution
of vital petroleum and natural gas could be hampered if electronically 
controlled pipelines malfunction."
   Of course, there is much more to the Y2K problem. Even if power
plants are ready for the rollover (which most of them are not at the
moment), will they be able to get fuel? The railroads aren't compliant.
If they can get the fuel, how will they pay for it? No major bank is
compliant as of this date, even though some of them have spent hundreds
of millions of dollars on it in a multiyear project.
  To conclude: Mr. Browne, however fine a person, a political candidate,
and an economist he may be, does a disservice to the public with this
article. Our citizenry need to be awakened to the seriousness of the
problem rather than put to sleep in the name of the free market.

Steve Heller
Return to my main page