The answers to these questions and others like them are not answerable with a yes or a
no. For one thing, the "leading experts" in the computer world cannot agree that there is even a
problem, let alone discuss the magnitude upon which it will impact society and the business
world. CNN correspondant Jed Duvall illustrates another possible "problem" scenario. Suppose
an individual on the East Coast, at 2 minutes after midnight in New York City on January 1,
2000 decides to mark the year and the century by calling a friend in California, where because of
the time zone difference, it is still 1999. With the current configurations in the phone company
computers, the NewYorker will be billed from 00 to 99, a phone call some 99 years long!!! (p. 1)
What if you deposit $100 into a savings account that pays 5% interest annually. The
following year you decide to close your account. The bank computer figures your $100 was
there for one year at 5% interest, so you get $105 back, simple enough. What happens though, if
you don't take your money out before the year 2000? The computer will re-do the calculation
exactly the same way. Your money was in the bank from '95 to '00. That's '00 minus '95, which
equals a negative 95 (-95). That's -95 years at 5% interest. That's a little bit more than $10,000,
and because of the minus sign, it's going to subtract that amount from your account. You now
owe the bank $9,900. Do I have your attention yet??!!
There is no industry that is immune to this problem, it is a cross-platform problem. This
is a problem that will affect PCs, minis, and mainframes. There are no "quick fixes" or what
everyone refers to as the "Silver Bullet." The Silver Bullet is the terminology used to represent
the creation of an automatic fix for the Yk2 problem. There are two major problems with this
philosophy. First, there are too many variables from hardware to software of different types to
think that a "cure-all" can be found that will create an "across-the-board" type of fix. Secondly,
the mentality of the general population that there is such a "fix" or that one can be created rather
quickly and easily, is creating situations where people are putting off addressing the problem due
to reliance on the "cure-all." The " . . . sure someone will fix it." type attitude pervades the
industry and the population, making this problem more serious than it already is. (Jager, p. 1)
People actually think that there is a program that you can start running on Friday night . . .
everybody goes home, and Monday morning the problem has been fixed. Nobody has to do
anything else, the Yk2 problem poses no more threat, it has been solved. To quote Peter de
Jager,
"Such a tool, would be wonderful.
Such a tool, would be worth Billions of dollars.
Such a tool, is a na ve pipe dream.
Could someone come close? Not very . . .
Could something reduce this problem by 90%? I don't believe so.
Could it reduce the problem by 50%? Possibly . . . but I still don't believe so.
Could it reduce the workload by 30%? Quite likely."
(p. 2)
Tools are available, but are only tools, not cures or quick fixes.
How will this affect society and the industry in 2000? How stable will software design
companies be as more and more competitors offer huge "incentives" for people to "jump ship"
and come work for them on their problems!? Cash flow problems will put people out of
business. Computer programmers will make big bucks from now until 2000, as demand
increases for their expertise. What about liability issues that arise because company "A" reneged
on a deal because of a computer glitch. Sue! Sue! Sue! What about ATM lockups, or credit card
failures, medical emergencies, downed phone systems. This is a wide spread scenario because
the Yk2 problem will affect all these elements and more.
As is obvious, the dimensions to this challenge are apparent. Given society's reliance on
computers, the failure of the systems to operate properly can mean anything from minor
inconveniences to major problems: Licenses and permits not issued, payroll and social service
checks not cut, personnel, medical and academic records malfunctioning, errors in banking and
finance, accounts not paid or received, inventory not maintained, weapon systems
malfunctioning (shudder!), constituent services not provided, and so on, and so on, and so on.
Still think you'll be unaffected . . . highly unlikely. This problem will affect computations which
calculate age, sort by date, compare dates, or perform some other type of specialized task. The
Gartner Group has made the following approximations:
At $450 to $600 per affected computer program, it is estimated that a medium size company will
spend from $3.6 to $4.2 million to make the software conversion. The cost per line of code is
estimated to be $.80 to $1. VIASOFT has seen program conversion cost rise to $572 to $1,204.
ANDERSEN CONSULTING estimates that it will take them more than 12,000 working days to
correct its existing applications. YELLOW CORPORATION estimates it will spend
approximately 10,000 working days to make the change. Estimates for the correction of this
problem in the United States alone is upward of $50 to $75 Billion dollars.
(ITAA, p. 1)
Is it possible to eliminate the problem? Probably not, but we can make the transition
much smoother with cooperation and the right approach. Companies and government agencies
must understand the nature of the problem. Unfortunately, the spending you find for new
software development will not be found in Yk2 research. Ignoring the obvious is not the way to
approach this problem. To assume that the problem will be corrected when the system is
replaced can be a costly misjudgment. Priorities change, development schedules slip, and
system components will be reused, causing the problem to be even more widespread.
Correcting the situation may not be so difficult as it will be time consuming. For
instance, the Social Security Administration estimates that it will spend 300 man-years finding
and correcting these date references in their information systems - systems representing a total of
30 million lines of code. (ITAA, p. 3) Common sense dictates that a comprehensive conversion
plan be developed to address the more immediate functions of an organization (such as invoices,
pay benefits, collect taxes, or other critical organization functions), and continue from there to
finish addressing the less critical aspects of operation. Some of the automated tools may help to
promote the "repair" of the systems, such as in:
* line by line impact analysis of all date references within a system, both in terms of data and
procedures;
* project cost estimating and modeling;
* identification and listing of affected locations;
* editing support to make the actual changes required;
* change management;
* and testing to verify and validate the changed system.
(ITAA, p. 3)
Clock simulators can run a system with a simulated clock date and can use applications that
append or produce errors when the year 2000 arrives while date finders search across
applications on specific date criteria, and browsers can help users perform large volume code
inspection. As good as all these "automated tools" are, there are NO "Silver Bullets" out there.
There are no quick fixes. It will take old fashioned work-hours by personnel in order to make
this "rollover" smooth and efficient.
Another area to look at are the implications for public health information. Public health
information and surveillance at all levels of local, state, federal, and international public health
are especially sensitive to and dependent upon dates for epidemiological (study of disease
occurrence, location, and duration) and health statistics reasons. The date of events, duration
between events, and other calculations such as age of people are core epidemiologic and health
statistic requirements. (Seligman, p. 1) Along with this, public health authorities are usually
dependent upon the primary data providers such as physician practices, laboratories, hospitals,
managed care organizations, and out-patient centers etc., as the source for original data upon
which public health decisions are based. The CDC (Centers for Disease Control and Prevention)
for example, maintains over 100 public health surveillance systems all of which are dependent
upon external sources of data. (Issa, p. 5) This basically means that it is not going to be
sufficient to make the internal systems compliant to the year 2000 in order to address all of the
ramifications of this issue. To illustrate this point, consider the following scenario: in April
2000, a hospital sends an electronic surveillance record to the local or state health department
reporting the death of an individual who was born in the year "00"; is this going to be a case of
infant mortality or a geriatric case??