I.
The size and nature of the Year 2000 problem
(1)
a)
Difficulties in estimating the size of the problem
Nobody has been able to pinpoint exactly how
widespread the year 2000 problem is. It is estimated,
however, that there are one trillion lines of
date dependent COBOL code and 25 billion embedded
chips worldwide. Of the embedded chips, 2 to
5 per cent are estimated to have the problem.
It is not easy to identify which lines of code,
and especially which chips are faulty. Many
analysts and software vendors are ready to admit
today that the problem is much worse than they
initially thought. For instance Microsoft, after
first denying and then maintaining silence about
the existence of the year 2000 problem, recently
admitted that its software products were not
fully compliant(2),
and opened a year 2000 Web site (http://www.microsoft.com/year2000/).
There is very little information available
about the prevalence of the problem in Asia
and the Pacific in general. A few quotes from
a handful of countries do not permit an assessment
of the overall picture. Awareness seems to be
high and work well under way in Australia, Hong
Kong (China), New Zealand and Singapore. On
the other hand, Japan, with a very large pool
of mainframes and custom-made software, has
according to foreign analysts made a very slow
start. Industry experts quote that the state
of (achieving) compliance in Japan is one of
the most worrying global year 2000 issues because
of the country's importance in global markets
(trade and finance). Knowing the region's aggregate
exposure would hardly make a difference in resolving
the problem, but might help in contingency planning
and risk analysis.
Looking at overall year 2000 budgets, private
enterprises seem to be taking the problem much
more seriously than public sector organizations
of the same size(3).
Globally, the resolution of the year 2000 problem
is estimated to cost $US one trillion, or significantly
more than calculated only a few months ago.
(b) The
nature of the yea 2000 problem
Affects all sectors. The year 2000 problem
affects computer hardware, custom-made and packaged
software, operating systems, telecommunications,
and of course data themselves. It can shut down
power plants and grids (including their backup
systems(4)),
prevent telecommunications, keep financial institutions
closed for extended periods, disturb traffic
on land, in the air and sea. Buildings can be
left without heating or air conditioning, and
lifts may stop(5).
Life support systems and other bio-mechanical
devices in hospitals can malfunction, and automated
industry processes can be halted. Military equipment
may malfunction. The list is long and covers
practically all economic sectors.
Invisible today and unpredictable.
It is estimated that the majority of modern
software and equipment would not be seriously
affected by the year 2000 problem. Unfortunately,
nobody can tell without testing which of the
systems will function correctly and which will
fail. Even software and equipment manufacturers
have been reluctant to give full guarantees
for their own products.
Elusive. Year 2000 system failures have
different impacts depending on the service or
process halted or made erroneous. The detection
and fixing of the problem would be fairly simple
if the failures were limited to a single point
per system. However, many of the year 2000 problems
are likely to be multiple point failures, requiring
complicated identification and testing methods.
Some software, claimed to be year 2000 compliant
by the supplier, have failed rigorous testing
done by third parties. Only fully-verified systems
can be claimed to be compliant.
Nothing new, therefore should have been
routinely dealt with. It is rather surprising
how large a problem year 2000 conformity is,
and how late the world is acting upon it, considering
that it is a result of a conscious, rational,
and industry-wide decision to save storage and
processing capacity(6).
The year 2000 problem was encountered for the
first time already decades ago in systems looking
30 years ahead. Moreover, analogous computing
difficulties have occurred in the past, and
will occur in the future, with other dates(7).
Time-consuming and costly to fix. Although
the remedies are often technically simple, it
can be time consuming to fix all the bugs in
a major system. Based on a survey conducted
in the United Kingdom, small organizations (less
than150 employees) were estimated to require
on average 18 months to become compliant, while
medium (150-500 employees) and large (more than
500 employees) organizations would take 24 and
30 months respectively. It is not unusual, especially
in organizations with a long history of mainframe
operations, for applications to be poorly documented,
for changes to be more so, and for the staff
that did the coding to be no longer available.
Often no alternative fixes. It may
be impossible to fix some hardware, which leaves
replacement as the only option. For embedded
microchips, there are no alternatives to replacement,
other than bearing with the nuisance caused
by malfunction. Embedded chips are (today in
the United States) considered to pose a greater
national risk than the computer applications
and hardware.
Requires spending without increases in
productivity. Achievement of year 2000 compliance
increases productivity only very little, and
then only because better replacements are installed.
Also, long term productivity can suffer as resources
are used for intensive year 2000 efforts instead
of new IT development (which is deferred).
A matter of urgency. The bottom line
is that there is very little time left for awareness
campaigns and overtures. Talking about the problem
is easy, but working on the solution can be
difficult and expensive. Year 2000 projects
must be started immediately, and should not
be delayed by red tape or resource disputes.
Testing is crucial, difficult and demanding.
Year 2000 testing has proved more difficult
and more time consuming than initially thought.
Most organizations spend more than half of the
time of Y2K problem solving in testing, and
75 per cent is not uncommon. Perhaps the greatest
difficulty is that testing can seldom be done
with a live system. Testing is required to identify
the problem and it must be repeated after the
fixes are in place. At both ends, testing involves
checking the four compliance criteria (see page
10). Especially critical, and laborious, is
the future date testing, which requires that
data are aged to simulate expected real data
inputs (transactions, orders, etc.).
There is no shortage of material about the
year 2000 problem and its solutions on the Internet.
For instance, the Australian national year 2000
site (http://www.y2k.gov.au)
demonstrates a systematic approach by the Government.
It emphasizes, among other things, the involvement
of top management echelons, systematic scheduling,
commitment and progress reporting.
The fact that it is almost impossible to get
insurance against year 2000 damage illustrates
the uncertainty(8)
involved in the year 2000 problem. Those few
quoted in the United States, charge annual premiums
between 25 and 33 per cent of the face value
of the policy. To issue a policy in the first
place, insurers want concrete evidence that
the organization, its suppliers, and major clients
have already undertaken substantive year 2000
remediation efforts. In other words, if no organized
effort has been made, no policy can be obtained
at any price. The minimum premium was quoted
at $325,000 with a $500,000 deductible, and
no new policies will be issued after early 1999.
(c) Doomsday
scenarios in perspective
Opinions about the impact of the year 2000
problem vary broadly. Almost daily one can read
that the scale of the problem is grossly exaggerated
and that its impact on individuals would not
go much beyond minor annoyances. On the other
hand, the nervousness is clearly growing among
the industry analysts and managers. That seems
a logical outcome: In many organizations laborious
investigations and inventories have just been
completed and the abundance of work has been
revealed. Achieving year 2000 compliance may
not be as easy as previously thought, which
gives a reason to believe that collaborators
are probably in the same position.
An all-round discussion is under way on the
Internet forums (Web, and Usenet group comp.software.year-2000)
about the impact the year 2000 failures are
going to have on various national economies
and the global economy. The following are some
of the frequently expressed opinions. While
many major multinational conglomerates will
be able to complete their conversion work in
time, some bugs will go undetected. On the other
hand, some enterprises will fail to achieve
compliance in time. The key question is whether
the number of failures will be so significant
that (i) it causes a major domino effect that
paralyzes economies, and (ii) makes people lose
trust in the capability of society and its players
(banking systems, logistical and delivery systems,
electricity supply, etc.) to recover quickly.
Many businesses are expected to be closed for
a good part of December 1999 because of the
frantic last minute efforts to get ready, or
because of effort to minimize the impact on
their own business. That alone would make a
significant dent in 1999 output(9).
Doomsday sayers also predict runs on banks'
cash reserves, which stand at around 6 per cent
of the stated deposits. A major stock market
decline is also being predicted for the months
well before the year 2000. Some predict even
that public panic is inevitable towards the
end of 1999.
So, how to react as an individual citizen?
What is a reasonable expectation for the change
of the century under current knowledge? Plenty
of critical work is under way, especially in
technologically advanced countries, and will
have been completed in 1999. However, we will
probably see delays and mistakes in some financial
transactions. Some businesses will have to shorten
opening hours, some will need to stay closed
until emergency fixes have been put in place.
Some daily goods could be temporarily in short
supply. There could be local blackouts and brownouts
(low voltage), but most of them will not be
very long (where the power grid is well-established).
Telephone calls may not be possible to all destinations,
and the fixing of telephone systems may take
some time. Some systems programmed to follow
working weeks (e.g. backup systems) will continue
to operate in a weekday mode on Saturday, 1
January 2000 and switch to the weekend mode
only on Thursday, 6 January 2000 (1 January
1900 was a Monday).
Eventually, life will return to normal as the
remaining year 2000 problems are fixed and faulty
systems are replaced. However, for some systems
and for some organizations, the problem may
be fatal and force them out of business. Therefore,
if anybody needs to panic, it is the management
of organizations which have not even started
to address their own situation.
Of mission-critical systems that must be fixed,
the smallest ESCAP member and associate member
economies may have none, other small ones may
have tens, a large number of countries have
hundreds, and the large or technologically advanced
countries have thousands. Those numbers are
of course much higher if pure PC applications
are included. It is the last opportunity to
start working on the solutions. The noncompetitive
nature of year 2000 problem offers opportunities
for cooperation between various organizations,
including in the private sector.
II. SELECTED
CRITICAL ISSUES
(a) Embedded
chips
The range of equipment and systems using embedded
systems is immense, and includes such domains
as manufacturing and process control systems,
building management, communications, office
facilities, and home electronics. While the
majority of embedded systems do not involve
time and date in their operations, it is not
always obvious which are at risk. For instance,
systems may have unused date and time functions
that could still cause failures. All equipment
and systems which contain embedded systems must
therefore be considered at risk until proved
otherwise (see for instance http://www.iee.org.uk/2000risk/updates/update02.htm#apndx_k).
The Gartner Group has warned that at least
10 per cent of mission-critical systems in health
care organizations will fail because of the
year 2000 problem. An inventory by a major hospital
in the United States found out that three out
of every ten pieces of biomedical equipment
would fail. Hewlett-Packard, a major medical
equipment supplier, has admitted that some of
its products will not work without a fix. Some
of the noncompliant equipment may have been
purchased as recently as in 1995.
Noncompliant chips may cause shutdowns of machinery
or equipment. While it will be possible to restart
many of them, some might require manufacturers'
service to replace the noncompliant chips or
upgrade their erasable programmable read-only
memories (EPROM).
Modern cars have 10-80 embedded chips, especially
in electronic engine management systems and
convenience systems. However, the manufacturers
are giving assurances that even the most advanced
cars would not be affected as the chips measure
time intervals, e.g. time passed since the previous
service rather than absolute dates. A more serious
impact to the car owner is thought to come from
the noncompliance of invoice payment, ordering
and other logistic systems of dealers and service
points.
(b)
Electricity supply
Electricity supply is the most crucial infrastructure
service that has a relatively high exposure
to the year 2000 problem. Power plants and grids
contain a large number of embedded systems,
and many are associated with dates. Power transmission
and distribution are highly automated and controlled
by computers. Nuclear plants have in addition
multiple safety systems which could force a
shut down if suspicious date values are recorded.
Should power generation or the grid fail, buildings
would be left without heating or airconditioning,
water supply would cease, and telephones would
most likely stop operating. It seems that many
electricity generating and distribution companies
(and authorities) have taken the year 2000 problem
seriously by starting their preparations relatively
early compared to many other sectors.
III.
THE YEAR 2000 PROBLEM IN GLOBAL SYSTEMS
The secretariat would have preferred to present
a picture about general year 2000 exposure and
preparedness in the region. Unfortunately, information
is available from only a few countries and it
was too patchy for any meaningful quantitative
analysis(10).
Instead, selected global systems are reviewed
with regard to the prevalence and impact of
the problem and the state of remediation work.
(a)
Financial institutions and financial markets
International (and national) financial transactions
today depend on computers and networks to the
extent that it is impossible to revert to manual
operations if their computers or communications
go down. Global financial transactions amount
to several trillions of dollars a day.
The year 2000 problem poses a very significant
risk of disruption to the operations of financial
institutions and financial markets. Therefore,
financial institutions worldwide are committing
large resources to become year 2000 compliant
in time. Apart from the workability of their
own internal systems, they must address external
systems with other financial institutions, clients,
counterparts and information vendors. Awareness
of the seriousness and scope of the problem
is in general high, but varies considerably
across markets and institutions worldwide. One
of the largest challenges is the compliance
of interbank or inter broker/dealer transactions
which cannot be conveniently tested because
of different individual remediation schedules
and because of the requirement of not disrupting
normal international transactions. It is therefore
not surprising that many national regulatory
authorities supervising and monitoring banking
and securities trade have taken a strong interest
in achieving sectoral compliance. Those authorities
are typically providing procedural recommendations,
committing their members to deadlines, demanding
progress reports, and making the year 2000 inspections
part of their regular examination cycles.
Since the problem could be foreseen a long
time ago and can invariable be fixed given sufficient
time, year 2000 failures make an ideal target
for law suits. Therefore, an audit of legal
implications should be a standard part of preparations
and contingency planning in any organization
supplying goods or providing services. Law makers
have imposed especially strict liability on
financial institutions, which in many circumstances
can be held liable even if they are not at fault.
For instance in the United States, banks are
liable for checks mistakenly bounced by them,
and for consequential damages suffered by the
clients.
International efforts in achieving year 2000
compliance in the financial sector are being
coordinated by the Joint year 2000 Council,
constituted by representatives of the Basle
Committee on Banking Supervision, the Committee
on Payment and Settlement Systems (CPSS), the
International Association of Insurance Supervisors
(IAIS), and the International Organization of
Securities Commissions (IOSCO). The Bank for
International Settlements (BIS) provides the
secretariat for the Council. (See http://www.bis.org/press/p980519.htm)
A paper by the Basle Committee on Banking Supervision,
"The year 2000 - A Challenge for Financial Institutions
and Bank Supervisors" (http://www.bis.org/publ/bcbs31.htm)
provides an overview of issues in the banking
sector, especially emphasizing the role of banking
supervisors' in promoting awareness, establishing
targets and benchmarks for the industry, and
even in using proactive supervisory pressures.
A document by the Bank of England, "Financial
Sector Preparations for the year 2000" (http://www.bankofengland.co.uk/y2t0598.pdf)
describes the multitude of issues that must
be addressed at national level to achieve compliance
in the finance sector
There is also another global coordinating body,
the Global 2000 Co-ordinating Group, formed
by several members of the global financial community,
to facilitate efforts by the global financial
community to improve the readiness of institutions
to meet the challenges created by the year 2000
(http://www.global2k.com)
Following the instructions from the Bank of
International Settlements, the Bank of Thailand
has instructed 15 local commercial banks to
tackle the year 2000 problem. The deadline for
running fully compliant systems is mid-1999.
That would allow a few months live testing before
the turn of the century. Banks not following
the established schedule would be barred from
dealing with other banks around the world. It
is obvious that such blacklisting would make
their banking operations very difficult, and
worse, could have a domino effect among enterprises
throughout the country, and beyond. The Bank
of Thailand says that it will try to fix its
own systems by the end of 1998 and use the first
six months of 1999 for testing.
(b)
Telecommunications
In the telecommunication sector, the Inter
Carrier Testing Subgroup of the International
Telecommunication Union (ITU) year 2000 Task
Force has identified the key risks as relating
to inter-carrier compliance(11)
(http://www.itu.org/y2k/minutes1.htm).
The Subgroup will attempt to produce a definitive
set of statements that will assist the ITU,
its members and their customers to establish
what is practicable and recommended when seeking
end-to-end assurances for year 2000 compliance.
The Subgroup was particularly concerned over
the lack of knowledge on the year 2000 issue
in developing countries, as they buy new technology
but do not often have the skills even for normal
maintenance. The interconnection between new
and old technology is also a problem, and not
only in developing countries. Issues related
to radio based communications also needed to
be addressed, especially in the Pacific.
(c)
Internet
The year 2000 Working Group of the Internet
Engineering Task Force (IETF) conducted an investigation
into the year 2000 problem as it regards Internet
related protocols in the following areas: "Autoconfiguration"
, "Directory Services", "Disk Sharing", "Games
and Chat (IRC)" ,"Information Services and File
Transfer (e.g. FTP, HTTP)", "Network and Transport
Layer", "Electronic Mail (SMTP, IMAP, POP, MIME,
X.400 to SMTP interaction)", "Network Time Protocol",
"Name Serving (DNS)", "Network Management",
"News", "Real Time Services", "Routing", "Security",
and "Virtual Terminal". (See details in draft
http://www.ietf.org/internet-drafts/draft-ietf-2000-issue-02.txt
of March 1998).
Although investigations are still not complete,
the Working Group found in general little reason
for concern with regards to the functionality
of the protocols. A few minor cases of older
implementations still using two digit years
(e.g. RFC 850) were discovered, but almost all
Internet protocols were given a clean bill of
health. Several cases of ''period'' problems
were discussed where a time field would ''roll
over'' as the size of field was reached(12).
Those limitations would be removed in new revisions.
The compliance of the Internet protocols does
not mean that individual users, or their Internet
service providers, could not experience serious
problems with their hardware and software respectively
(other than protocols), or that there could
not be other year 2000 problems with network
connections (e.g. in LANs or telecommunications).
However, in the light of the year 2000 compliance
of the Internet protocols (which are crucial
for data interchange) it is currently thought
that key parts of the Internet would remain
operational at the turn of the century.
(d)
Global Positioning System - a dual rollover
problem
According to the United States Department of
Defense, the Global Positioning System (GPS)
would have malfunctioned due to year 2000 code
problems without the repairs that are under
way (http://208.228.76.74/gcn/1997/april14/dod.htm).
Without additional fixes, the GPS would also
suffer from another date related rollover problem.
The "end of week problem" would occur 132 days
prior to year 2000 (at 0000 hrs UTC, 22 August
1999). It arises from the fact that the GPS
time count, which started at midnight 5/6 January
1980 and which is counting weeks, will reach
the ceiling number 1023 (2^10=1024) and turn
to 0000. Many GPS receivers in ships and air
crafts would then operate as if it was 6 January
1980, which could affect the accuracy of navigation.
It is reported that all generations of the
24 GPS satellites in orbit are unaffected by
the two problems. Some satellite support system
software are not year 2000 compliant and are
scheduled to be replaced or repaired. The operational
GPS control system operates on a legacy mainframe
system and is affected by the year 2000 problem.
The year 2000 repairs are expected to be completed
by the end of 1998. Some handheld GPS receivers,
mostly older models, are not year 2000 compliant.
As there is no way to test the receivers, users
are encouraged to contact the manufacturer to
determine if the two problems occur in their
devices and affect navigation. (See a list of
manufacturers at http://www.navcen.uscg.mil/gps/geninfo/y2k/gpsmanufacturers/manufacturers.html,
and some compliant receivers at http://www.laafb.af.mil/SMC/CZ/homepage/y2000/comply.htm).
IV.
ACHIEVING COMPLIANCE
Exposure to the year 2000 problem depends of
course on the existence of software, hardware
and embedded systems that use two-digit dates.
The following gives indications about the kind
of organizations that have low and high exposure
to the year 2000 problem.
| Factors that reduce (but do not
eliminate) Y2K problem exposure |
Indications of Y2K problem exposure |
| Small organization |
Large organization |
| Low rate of computerization |
IT intensive working environment |
| IT development started in full force less
than five years ago |
Long computerization history, application
development started in the 1980s or earlier |
| Low rate of electronic technology in buildings |
Modern and convenient office environment |
| PC-based computing |
Use of mainframe computers, client-server
technology in general (studies show that
most new mainframes run old applications) |
| Use of latest operating systems |
Old operating systems; use of several
operating systems |
| Use of latest off-the-shelf software |
Custom-made applications |
| New PCs and LAN equipment |
Large stock of old PCs |
| Low dependence on telecommunications |
Business based on telecommmunications |
| Low rate of electronic transactions with
collaborators and clients |
Electronic exchange of data and information |
| Systems are of recent development,
and developed with latest development tools
|
Applications are old and poorly documented |
| Staff responsible for development of systems
are no longer with the organization |
| Haphazard and poorly documented data backup
strategies and systems |
Falling in the left column does not mean that
the organization could not encounter year 2000
problems. The chances for the occurrence of
the year 2000 problem are many. To be able to
say with a high probability that the component
is not affected by dates, all of the following
criteria should be met (taken from http://www.iee.org.uk/2000risk/updates/update02.htm):
- There is no real-time clock
hardware.
- Timers do not use a difference
in dates to calculate a time.
- There is no battery backing
for the processor or memory. This point should
not be taken as sufficient evidence in itself
as some chips maintain information (and potentially
therefore date information) even though there
is no visible means of power.
- There is no access to internal
or external non-volatile devices, such as
disks and tapes.
- The date is not available
(known) at the Operating System layer.
- The application language
contains no constructs or libraries that use
or manipulate dates.
- There is no external date
interface from clocks or over communications
lines.
- The operator never sets
a date in the system.
- The operator never sees
any dates (on input or output).
If the last three conditions (g, h, i, the
detection of which do not require deeper technological
knowledge) are true and if the system concerned
has also low potential impact on the operations
of the organization, the risk from that system
may be deemed low, and further investigations
could be given a low priority. However, as the
year 2000 problem is often invisible (especially
in embedded systems), systems whose failures
have potentially greater impact would need to
be assigned a high testing priority.
Year 2000 conformity means that neither the
performance nor the functionality is affected
by dates prior to, during and after the year
2000. Full conformity means that
No value for (current) date will cause any
interruption in operation.
Date-based functionality must behave consistently
for dates prior to, during and after the year
2000.
In all interfaces and data storage, the century
in any date must be specified either explicitly
or by unambiguous algorithms or inferencing
rules.
Year 2000 must be recognised as a leap year
(29/2/2000 and 366 days in the year 2000).
(Source: British Standards Institution Committee
BDD/1/-/3, DISC PD2000-1: A definition of year
2000 conformity requirements, http://www.iee.org.uk/2000risk/guide/year2k98.htm)
V.
THE ROLE OF ESCAP
The tenth session of the Working Group of Statistical
Experts (WGSE) in November 1997 was the first
ESCAP meeting to consider the implications of
the year 2000 problem in the region. After hearing
threatening scenarios on the one hand, and large-scale
resource mobilization by some advanced statistical
offices on the other, the Working Group recommended
taking the topic up at a higher level, the annual
Commission session (held in April 1998). The
current Workshop on the year 2000 Problem in
Computers and Strategic Issues for National
Statistical Offices is another follow-up event
to the WGSE meeting. The recommendations of
the WGSE and the 54th session of
the Commission are extracted below for easy
reference and possible recapitulation and refinement
by the Workshop.
The ESCAP secretariat has been creating awareness
through the Government Computerization Newsletter
and its Web site. However, it does not have
operational resources to assist members and
associate members in a more tangible way, such
as by providing advisory services or responding
to individual technological questions.
(a)
Working Group of Statistical Experts
The Working Group of Statistical Experts, a
subsidiary body of the ESCAP Committee on Statistics,
recommended at its tenth session in November
1997 that NSOs identify the implications of
year 2000 failures in of any of their systems
in order to prioritize which of the mission-critical
applications should be fixed first. Having reviewed
the serious technological and managerial challenges
that all countries in the region were facing
within a very short time, and having compared
them to scarce resources and low awareness,
especially in developing countries, the Working
Group requested:
- The secretariat and the
Bureau of the Committee on Statistics to create
awareness through the Commission session in
April 1998 that the year 2000 problem posed
a real, serious and potentially economically
hurting threat to the governments in the region
and that they needed to allocate resources
urgently to tackle the problem;
- The chairperson, secretariat
and the members of the United Nations Statistical
Commission from the region to raise the issue
in the forthcoming session;
- The secretariat to create
awareness of the year 2000 problem in the
countries of the region by compiling and disseminating
information through its publications and web
site; such information should use non-technical
language and be disseminated widely;
- The secretariat and the
NSOs to facilitate the sharing of experience
in the region, especially from the governments
and NSOs that had tackled the problem with
some success and comprehensiveness;
- The secretariat to approach
the donor on the possibility of including
the year 2000 issue on the agenda of the planned
seminar on IT applications, and hold it in
early 1998;
- The secretariat to investigate
if meetings on the year 2000 issue could be
held soon outside the standard project funding
cycle;
- SIAP to investigate if it
could organize a training event in early 1998
on the year 2000 issue.
The Working Group cautioned NSOs not to wait
for information on other countries' experiences,
since those countries were far from completing
their own solutions.
(b)
Fifty-fourth session of the Commission
The Commission expressed, in its fifty-fourth
session in April 1998, a deep concern about
the predicted disruptions that the year 2000
problem in computers and embedded chips was
likely to cause at the national, regional and
global levels. Noting the slow start made by
many countries of the region in tackling the
problem, it urged all governments to make its
resolution a high priority. The Commission recognized
that the problem was not restricted to statistics,
but also affected infrastructure services such
as electricity supply and telecommunications,
as well as banking and other systems. The Commission
emphasized that it was the responsibility of
top level management to initiate organization-wide
action to address the issue. For identification
and resolution of the problem, the Commission
recommended the use of multidisciplinary teams
that periodically reported on progress to high-level
management.
As an immediate measure, the Commission recommended
that organizations demand guarantees from suppliers
that all new software and equipment were year
2000 compliant. The Commission advised all organizations
to make contingency plans in case of failure
of their own systems or of external or foreign
systems that they were increasingly dependent
on. Given the urgency of the situation, the
impending high work volume in fixing existing
systems meant that mission-critical applications
had to be given the highest priority. The Commission
warned that any delays were likely to increase
the modification cost and make the timely resolution
of the problem very difficult, as the required
skills were already in short supply.
The Commission endorsed the recommendations
of the Working Group of Statistical Experts
in regard to the year 2000 problem. While
recognizing that the problem could only be
solved at the level of each organization,
the Commission encouraged all members to share
their experiences in resolving the year 2000
problem and asked the secretariat to facilitate
such regional cooperation.
(c)
Secretariat's own compliance
The ESCAP secretariat recognizes that its own
systems are not immune from the year 2000 problem,
and has therefore been addressing the issue
across the spectrum of its computer hardware,
software and databases, telephone systems, lifts,
building automation systems, card-key systems,
etc. A year 2000 coordinator has been designated
from among the staff of the Electronic Services
Section, and technical information about testing
PCs has been distributed to the divisions.
The United Nations as a whole is in transition
to a comprehensive system-wide management information
system (United Nations Integrated Management
information System, IMIS) for accounts, payroll,
personnel, etc., processing. The development
of IMIS is undertaken centrally at the United
Nations Headquarters, and they provide updates
as required for the already-installed software
at ESCAP and other major duty stations, and
advise on the need to upgrade the hardware and
operating systems for the host computers. The
software modules already installed and those
to be installed prior to the end of century
are fully year 2000 compatible. The hardware
on which IMIS runs is certified to be year 2000
compliant.
There is, however, some delay in the schedule
with regard to the installation of the next
major modules of IMIS, namely for payroll and
accounts, and ESCAP's legacy payroll and accounting
systems, which are known to bear the year 2000
problem, would continue in operation instead.
Their COBOL code would need to be fixed, which
work will be done by using in-house knowledge.
The only problem is that computer programming
staff are already performing many other assignments
as a result of the already very tight resource
situation in the ESCAP secretariat. Another
concern arises from the drastic cut in ESCAP's
1998 hardware and software replacement budget,
which will delay the replacement of noncompliant
equipment and software.
The ESCAP Statistical Information System, a
client-server system based on Sybase and Powerbuilder
is said by its developer to be fully compatible,
as it is fairly recent development. However,
it has not yet been tested.
Footnotes:
1. The
year 2000 problem in computer hardware and software
and embedded chips refers to their inability
to handle (read, process, write) year 2000 related
date information (year, week day, leap year)
correctly. The problem stems from the long-standing
programming practice of using only the two last
digits to represent year information.
2.
MS Windows 95 and early applications made for
that environment, such as MS Office 95 have
minor date related problems while the very latest
releases are fully compliant. Access 2.0, Word
5.0 for DOS, and Office v. 4.3 are not compliant.
3.
According to Cap Gemini, budgets dedicated to
resolving the year 2000 crisis in public sector
organizations [in the United Kingdom] were only
one fifth of those found in private organizations
of comparable size.
4. Modern
uninterruptible power supplies (UPS) contain
a lot of software that manages the power supply
process.
5. Apart
from containing programs controlling the movement,
lifts monitor the time passed since the last
maintenance. If that period surpasses a set
limit, the lift will be bought to the ground
floor or basement and halted. Similar logic
is used to ensure maintenance in many other
devices, including medical devices.
6.
That decision looks rational even today as the
total cost of fixing it is estimated to be less
than the benefits already gained by two-digit
coding.
7. See http://www.iee.org.uk/2000risk/updates/apxb0598.htm
for a long list of dates that could cause difficulties
in computers and embedded systems.
8. Note the
difference between uncertainty and risk. Insurances
are normally available against risks whose probability
can be estimated.
9. Monthly
economic time series will be in great demand
for 1999-2000, immediately as they become available,
and long afterwards.
10.
The Workshop will be in a position to make such
an analysis in one sector and perhaps give some
indications about governments' overall preparedness.
11. They
were the element management around the transport
layer, time synchronisation between networks,
billing and settlement systems, IPC interfaces
, C7 signalling, invoicing over the network
(GSM), calling cards, GSM, and international
freephone calls.
12.
In particular, there are several protocols,
which have 32 bit, signed integer representations
of the number of seconds since January 1, 1970
which will turn negative at Tuesday, 19 January
2038, 03:14:07 hrs GMT. |
About Us | Media Centre | Members | Programmes | Documents | Publications | Jobs | 
Bangkok, Thailand