/dq/media/agency_attachments/UPxQAOdkwhCk8EYzqyvs.png)
Follow UsCan you imagine a day when:
-
-
-
-
Yes, we may have to face a horrible
situation. Because from that day onward the entire computer system may become useless.
What exactly is
Y2K problem
The full form is as seen here:
Y means Year, 2 means 2, and K means Kilo. Which is otherwise known as century date change
(CDC) or millennium bug.
In the year 1968, when
computerization was being done for all business processing applications, a six-digit
format was standardized by an American federal information processing standard.
In the six-digit format, the frst
two digits signify the month, the next two digits signify date and the last two digits
signify year.
But why store
only six digits
During the sixties the memory was so limited, that it was considered wasteful to store a
century value universally as '19.' But the members of that standard committee failed to
imagine that after 32 years, billions of lines of written code would result into nothing.
The result is that on January 1,
2000, the last two digits become 00, and many computer systems will not be able to
distinguish between the 20th and the 21st century, and their dependent data will become
confusing.
Now consider January, 3, 2000 in
six-digit format, that is 03/01/00 on comparing it to 31/12/99. Any computer program will
treat 31/12/99 as the later date.
Moreover, the year 2000 is a leap
year. We know that a year if divisible by 4, is a leap year. However, century years are
leap year only if they are divisible by 400. Thus 1900 was not a leap year, but 2000 is a
leap year. So the software will not be able to handle the date February 29, 2000.
Time is slipping
away One of earliest occurrence of the year 2000 problem was in the seventies when banks
needed to perform calculation for 30-year loans. In 1990 it was 10-year certificate of
deposit.
This year it will be driver's
renewals. So the problem is widely spread.
Real business
risk
Y2K is not technical problem. It is business problem. The system that supports our
business is at a risk of failure or the date integrity may be lost. Therefore if the
problem is not solved, the business itself will be at risk. For example:
-
the year 1991, which matures in 15 years that is in the year 2005, the difference will be
85 years instead of 15 years which is very risky for the insurance company.
Things that make one mad!! Like, a 107-year old woman in Ahmedabad received a letter from
the local school welcoming her to the Ist standard.
-
collating report dated 00 prior to 99, or incorrect automatic date calculation on report
and latter.
-
the 100 year marker, social security or public assistance check not being issued, credit
cards being rejected or defense system failing.
Not a cobol
problem
Y2K problem is not just taking the Cobol source code available on the mainframes and
converting the date format there in 80% of the millennium bug lines with the firmware. You
may think that only older Cobol applications on primordial mainframes have Y2K problem.
Sorry, but you are wrong, your desktop computer, your applications, your databases, your
spreadsheets and your form letters could all have the same date problem.
In application software area, if
source code is available, at least 10-15% of the code needs to be changed but if source
code itself is not available for a large number of software, identifying date-related code
among the object code is a difficult task.
Today, almost every company depends
on huge software applications written mostly in Cobol. According to industry sources,
there are more than 180 billion lines of Cobol code currently in use, with thousand of
such sites all over the world. Since these codes use a two-digit date field, many
applications will either stop producing any results or start producing incorrect results
at the start of the new millennium.
Consequences
The Year 2000 will affect every government agency, business or individual who uses a
computer, even if you don't use a computer, you will still feel the impact when you
interact with anyone such as banking, insurance, credit card and accounting system. So
this millennium bug needs to be solved even if it costs billions of dollars.
A bank's computer does this simple
operation for calculating the interest on, say thirty-year mortgage. There is no problem
if the mortgage were issued in 1965 and become fully payable in 1995, but if the mortgage
issued in 1970, is to become payable in year 2000, it is a cause of worry. If the
application uses two-digit year format, the application will read the year 2000 as 00 and
do the necessary calculations accordingly. As a result, most of the applications will
calculate the number of year (-70).
Suppose one of the functions of an
accounting system is to provide a list of all those invoices that have been outstanding
for more than a month. If the new order is given on invoice date of 02/02/00 then the
accounting system will mark it up as an order that has been outstanding for hundred years.
Year 2000 problem may result in
mistakes in calculating interest or delay in payments. Consequently, automatic checks or
premium notices may expose a company to legal risk.
Computer architecture has been
changing the worldover, from mainframes to client server systems. Today, even as
organizations move to a client server set-up, they must solve the parallel problem of
dealing with two-digit date fields with networked systems, all your need is one program on
one node in an application system to reject the 2000 date, and the whole network is
useless, for that date transactions may not reconcile for days afterward.
Solution process
is costly
The total business cost of Y2K problem is 200 to 600 billion dollars. For solving this
problem four years ago people thought that we have a lot of time before the arrival of the
Year 2000 at that time to solve this problem we had to spend about $0.5 per line which is
increasing day by day. Today, for solutions we have to spend $4 per line, which is likely
to go up to almost $7 by 1999.
This is due to the large gap
between the demand and supply of skilled manpower and the huge work involved.
No easy way out
There is no magic silver bullet that can sweep one's system to detect and automatically
correct CDC-code problems, and ensure that the resulting changes will work. However, there
are many tools that can be used to assist the CDC-compliance effort, if used properly as
part of a comprehensive conversion approach.
Reprogramming code to differentiate
between the year 1900 and 2000 is actually a simple task that can be accomplished by any
competent programmer. The difficulty is that from code much of it is in poorly documented
legacy systems that must be searched to identify CDC problem.
There are many software whose
source code itself is not available because it has been destroyed to make sure that
company's private information is kept secret and its protection ensured.
Such programs have object files.
But it is impossible to read it even if an expert tries to read the object file and
correct the date, the correction of one line of object code would cost around $20. Hence,
the company cannot afford to correct the object file. Moreover, there would be need for
lot of experts.
Y2K solutions
To solve this problem, till some new solutions are not discovered, there are only some
ways:
-
expand the year field to four digits.
-
because January 1, 1972 also being a Saturday and 1972 is a leap year and use the bridging
software that will add/subtract 28 years from the date filed. If the first method is
adopted, even if crores of people spend their entire 24 hours repairing the data, still
the entire problem cannot be solved.
-
the base year. Then codes in program are checked to find out if the year is less than 60?
If the answer is 'yes,' then 2000 is added to the year, otherwise 1900 is added to the
year. The problem with this method is that it does not work with certain applications.
(For example, Life Insurance).
Y2K tools
For managing Y2K project and solving Y2K problems, a large number of tools have appeared
in the market over the past few years. These tools can be divided into following
categories:
-
of Y2K.
-
help the project managers.
-
inventory of the software that needs solutions for Y2K problem.
-
object code
-
that simulates the required date without affecting operating system operations.
-
control.
Indian scenario
Here in India, the NASSCOM has formed a Special Interest Group to target this problem and
$600 billion world market that it offers.
There are 50 or more member
companies in the SIG. For example:
-
-
tools, smart 2000 to address the problem.
-
Microsoft Corp for its solution.
-
Signature 2000 as an integrated life-cycle solution.
-
worth about Rs20 crore.
Opportunity
factor
The Y2K problem provides the unique opportunity for many developing countries rich in
skilled manpower. Nothing today is being discussed more than the rise of software industry
in India and solutions for the millennium bug.
Y2K problem has certainly generated
a lot of business for India and the country can greatly benefit from this opportunity.
There are many factors that favor
the software industry in India:
Quality: India may soon become the country to have the largest number of ISO-9000
certified software companies in the world.
Cost-effective solution: India is in a unique position to offer error-free, quick
and efficient solution for Y2K problems. Software industry in India has already earned the
reputation of providing cost-effective solutions as compared to the other competing
countries.
State-of-the-art technology: Sixty-five percent of Indian software companies are
using state-of-the-art technologies. The companies are rapidly adding to the usage of Y2K
tools, 4GLS, Object-Oriented Programming, software reengineering, client server computing
etc.
Off-Shore development: It involves using high-speed datacom links to enable
programmers based in India to use computers located anywhere in the world on a real-time,
online basis.
Myths surrounding
Y2K jobs
There is a major demand of computer experts in India to solve this problem. Experts are
encouraged to enter this field. On the contrary, people claim that this is needed only to
last up to year 2000. But what about after the year 2000?
•
projects also involve reengineering of codes written in a variety of languages, example,
RPG, PL/I, C and even C++.
•
tools run on latest platforms like Windows95, WindowsNT, OS/2 etc.
•
experience is not counted: Cobol is being extensively used on Windows, Unix platforms for
client server and web-based application as well. In 1998, a standard will be released for
object-oriented Cobol, putting it on a solid platform for a take off to the 21st century.
Mainframe knowledge will be useless after two years: Mainframes have made a come back in
the projects such as datawarehousing and digital libraries.
DIPA J PATEL,
SONAL M SHAH
and RAJANI PATEL.
Excerpted from
conference proceedings:
CSI'98, 'IT For The New Generation.'
Courtesy Tata McGraw-Hill.
Box:
Y2K Affects
Area What's Affected Problem Enterprise hardware Mainframes Hard wired to use two-digit
year Enterprise application Cobol code lists or records Records & programs assume on
computer cards two-digit years Desktop hardware Desktop computers PC BIOSES can't handle
years after 1999 Desktop software Databases, Spreadsheets Assume two-digit years
Applications & Word Give wrong year after 1999 processors Files Spreadsheets &
form letters