|
1. Background
and General Overview of the Present Situation
What is the Year
2000 problem ?
The Year 2000 problem, also popularly known
as the "millennium problem" or the "Year 2000
crisis or disaster", is a hot topic not only
among professional in the Information Technology
(IT) industry but also users of IT in all industries
ranging from the manufacturing to financial
and services industries, and in both the private
and public sectors. To the extent that statistical
services are developed based on computerized
systems and involve, in one way or another,
processing of "date" information, they are also
affected by the Year 2000 problem.
The Year 2000 problem, as many of us are aware,
has arisen from the past practice adopted by
IT professionals of referring to the year by
its last two digits (e.g. using "97" to represent
"1997") in designing computer systems. The two-digit
year approach has been used mainly to economise
on computer storage and processing power which
were relatively expensive in earlier years.
Computer systems handle dates (and therefore
"years") in a wide variety of ways, ranging
from basic processing such as sorting, comparison
and validation to complex calculations. Records
are sorted in chronological order; calculations
and comparisons are made based on people's age;
and forecasts of different kinds make extensive
use of dates. All such processing may fail or
produce erroneous results when the year 2000
is reached if the century part of the date is
missing. For example, in compiling statistics
on "age at first marriage", a person born in
1970 who marries in 2000 would be assumed by
the computer, using a "two-digit" year approach,
to marry at the age of -70 (i.e. 00 - 70) instead
of the age of 30 (i.e. 2000 - 1970).
General overview
of the present situation in the Census and Statistics
Department (C&SD) of Hong Kong
The Year 2000 problem is multifaceted
and is present at various levels within a computer
system, namely, the hardware, the operating
system, the system software, the application
software and application systems. Computer
application systems can, by and large, be categorized
into end-user systems and non end-user systems.
End-user systems are application systems developed
by end users using commonly available, user
friendly software packages. Non end-user systems
are often large scale systems developed by professional
computer analysts and programmers. The Year
2000 problem affects both end-user and non end-user
systems and in the former case the end user
would play a more prominent role in tackling
the problem.
In the Census and Statistics Department
(C&SD) of the Government of the Hong Kong
Special Administrative Region, a preliminary
assessment on the Year 2000 compliance of its
systems had been conducted. At the hardware
level, 80% of the personal computers (PCs) and
90% of the Local Area Network (LAN) servers
have been identified as non Year 2000 compliant. The operating systems and the system
software of the IBM mainframe and HP mid-range
machines have to be upgraded before they are
Year 2000 compliant while the VAX minicomputer
is not upgradable to Year 2000 compliance version.
As for application systems, the assessment indicates
that a majority of these systems will need to
be modified to survive the turn of the century.
A brief summary on the Year 2000 compliance
status and recommended approach for achieving
such compliance for application systems on the
IBM, VAX and HP is shown below.
Manpower Survey System
| Platform
| Application Systems
| Modification
needed for Year 2000 Compliance
| Approach
for Year 2000 Compliance
|
| IBM/MVS
| Central Register of Establishment
System
| Y
| R
|
| Government Revenue &
Expenditure System
| Y
| C
|
| Full Harmonized System
| Y
| C
|
| Frame Reporting Units
System
| Y
| C
|
| VAX/VMS
| Trade Verification Information
System
| Y
| C
|
| Quarterly Survey of Construction
Output
| Y
| C
|
| Wages & Payrolls System
| Y
| C
|
| Trade Consignment Coverage
Checking System
| Y
| R
|
| Notice Issuing
for Non-lodgement of Trade Declarations
| Y
| R
|
| HP/UX
| Register of Quarters System
| Y
| C
|
| Programme of Economic
Surveys System
| Y
| C
|
| General Household Survey
System
| N
| R
|
| Survey on Employment and
Vacancies System
| N
| R
|
|
C-CONVERSION
R-REDEVELOPMENT
|
2. A Cost-effective
Approach to Tackle the Problem
While there is no one best way to tackle
the problem, an effective strategy may be developed
for tackling the Year 2000 problem, comprising
the following elements: impact assessment, team
formation, utilization of automated tools, information
sharing and setting up a help desk.
Firstly, an impact assessment would be conducted
to identify potential application systems for
conversion, redevelopment or retirement; to
define the project scope for conversion and
redevelopment; and to set priority on application
systems for conversion and redevelopment.
Secondly, a technically strong project management
team would be set up. Its scope of duties would
include obtaining funding approval, evaluating,
recommending and procuring hardware and software
required for locating, fixing and testing of
defective codes, and implementing the necessary
changes for the Year 2000 change. Besides government
staff, experts from the private sector would
be employed on contract basis where the conversion
and/or redevelopment work is significant and
complex.
As far as possible, automated tools would
be employed for tackling the Year 2000 problem.
This is often more cost effective than developing
an in-house solution. Tools like "TransCentury"
could be useful for fixing the Year 2000 problem
and expediting the conversion of IBM/MVS Cobol
Programs. Similar products for other platforms
would also be sourced.
A commonly shared knowledge base, either in
the form of a bulletin board or a data base
would be set up to facilitate sharing of information
and reference material on Year 2000 including
planning and implementation guidelines, compliance
criteria, progress updates, etc. At C&SD,
such a data base has been set up within the
Lotus Notes Office Automation system.
Last but not least, a help desk would be set
up to provide technical support services on
the Year 2000 problem for all application systems.
This would be especially welcomed by end users
tackling the Year 2000 problem for systems under
their purview.
3. Implications
on End Users and the Importance of End User
Involvement
As mentioned above, end user application systems
are similarly affected by the Year 2000 problem
and the various areas outlined above are also
relevant for these applications.
A distinctive feature of end-user application
systems is that they are developed by end users
themselves and often involve greater subject
matter complexity. End users are thus most knowledgeable
on system design and other features and it is
therefore crucial that they play a major role
in tackling the Year 2000 problem. Their participation
and contribution towards impact assessment,
assessing and deciding whether the system should
be converted or redeveloped, implementing changes
and conducting acceptance tests is important
and effective. In some cases, end users could
actually assume the role of project leaders.
Besides end user participation, technical
support for tackling the Year 2000 problem is
also essential. Depending on the size of the
system, such support could range from consultation
provided by the Help Desk mentioned earlier
to having a full time computer analyst or programmer
on the project team. Where necessary, computer
staff in the private sector may be employed
on contract basis for technical support and
for implementing the changes.
4. Anticipated
Problems
As we prepare to tackle the Year 2000
problem, a number of problems can be anticipated.
First and foremost, the deadline cannot be
deferred. It is therefore pertinent that the
implementation schedule must be met through
close monitoring.
Secondly, besides modifying or redeveloping
application systems, success with meeting the
Year 2000 challenge also depends on the availability
of Year 2000 compliant hardware, system software,
operating systems and application software packages.
This in turns depends on the capability of the
various manufacturers to develop and make available
such products in time for users to convert or
redevelop the application systems. At present,
some Year 2000 compliant operating systems,
system software and application software are
still not available and, worse still, in some
cases even the availability dates have not yet
been announced. Application systems using such
hardware or software would face a higher risk
if Year 2000 compliant versions are made available
at a very late stage thus leaving even less
time for system conversion or redevelopment.
Thirdly, the appropriate financial authorities
must be alerted to as early as possible, and
convinced of, the significance of the Year 2000
problem and provide the necessary financial
resources to tackle the problem.
Fourthly, besides hardware and software tools,
the availability of experienced programmers
is another critical key for successfully tackling
the Year 2000 problem.
5. The Way Forward
To successfully tackle the Year 2000
problem, early planning, impact assessment and
project sizing is important. Where necessary,
industry experts should be consulted. Besides
seminars and discussion sessions, web sites
dedicated to the Year 2000 problem are available
and could be consulted to acquire greater exposure
and knowledge on the issue. A list of useful
web sites is provided in the Appendix. Moreover,
close liaison with hardware and software manufacturers
and vendors should be maintained to ensure that
up-to-date information on Year 2000 compliant
products is available. Lastly, provision should
be made for sufficient and competent manpower
resources and funding to meet the Year 2000
challenge.
- CENSUS & STATISTICS DEPARTMENT
- HONG KONG, CHINA
- November, 1997
Appendix
Some Useful Web
Sites on the Year 2000 Problem
|
