Sponsored by
INTRODUCTION
There are three basic ways to cover something up. The most blatant way is simply
to CENSOR it. By excluding a certain fact from news coverage, for example, the truth
simply does not get out in the open where people can hear it -- and act upon it.
The second way to cover something up is to OBSCURE it. By placing an important piece
of news on a back page, or hidden among a dozen other high-profile stories, or under
a tiny headline (or a misleading headline that disguises the import of the news
item), it is easy to keep the public from recognizing the significance of a piece
of news.
But the most subtle and sometimes most effective means of deception is not always
intentional. Sometimes, simply by remaining blind to the dangers posed by a particular
product, company, or person, the news media can unintentionally spread a comfortable
numbness, a fuzzy feeling of nonchalance, or even a mindless acceptance of something
that should rightly be exposed as a major potential hazard. It is thus a cover-up
of a hazard by choosing to IGNORE it that results in a whitewash of the facts, and
in a failure to properly warn the public of an impending danger. Is that not one
of the responsibilities that the media claims to fulfill, to be a warning sentinel
for society?
Probably the most blatant example of this third form of news-hiding is the computer
media's contemptible coverage of the Y2K issue. For ten years or more, the PC news
outlets such as PC Week, Info
World, and PC Magazine never even hinted
at a possible Y2K problem with the products they reviewed, recommended, and promoted.
It was not until the late 1990's that the issue finally began to see the light of
day -- when it was nearly too late to do anything about it. By the time 1998 rolled
around, most companies had standardized on a set of Microsoft products that was
neither Y2K-ready, nor was it readily verifiable for Y2K compliance. Since the computer
press had long been asleep at the wheel and had remained ignorant of this issue,
they could do no better than to pass on their ignorance to the trusting multitudes
of readers who put faith in their every word.
Yet even now, halfway through 1999, not one solitary mainstream computer magazine
has made a clear, direct, realistic recommendation as to how to escape the Y2K meltdown
that the vast majority of PC users are currently threatened by. Even now, the vast
majority of products that the PC press continues to review and recommend unconditionally,
simply cannot be ready for the year 2000. And what is the main reason for this situation?
It is because the software platforms these applications require -- the operating
systems themselves -- are not ready for Y2K and probably will not be ready.
Most of the concern about Y2K and its effect on PCs has focused exclusively on PC
hardware. Walk into the local computer store in your area and ask to see the Y2K
policy for any of the computers on sale. What will you read? A typical advertisement
will promise "Y2K Hardware Compliant" products. However, this is a little
bit like offering someone a cup of water and guaranteeing that the cup itself is
"100% poison-free." WHAT ABOUT THE CONTENTS OF THAT CUP?
At present, no "mainstream" consumer PC vendor that I am aware of will
guarantee that the entire product they put on the store shelves (hardware and software
together) is 100% Y2K-ready. Not a single vendor. As a wise old person once said,
you can sometimes find out more from what a person does NOT say than from what the
person DOES say. If the PC vendors really believe Microsoft products are 100% ready
for Y2K, why do they not say so in the advertisements? Why not a small statement
that simply points to Microsoft as the guarantor of the Y2K-readiness of the operating
systems and associated applications? Because both PC vendors and Microsoft know
full well that they are not in any position to make such a guarantee. Their consciences
would probably not bother them, but their legal departments would forbid it. Selling
junkware is one thing; selling junkware that is doomed by the sands of time is only
acceptable to these companies if they can get away with it by not drawing any attention
to it. Since the average PC buyer thinks the operating system is part of the hardware,
the average PC buyer is fully deceived by the guarantee of "Y2K hardware compliance."
It will only be later when the software defects begin to occur that the real education
will finally begin.
As for the handful of PC buyers who look beyond the surface, they are being deceived
by the guarantee of Y2K compliance that is sometimes affixed to a software application.
Believing that if the PC is "100% hardware compliant" and the application
being purchased is "Y2K ready," even the most astute buyer of a so-called
"mainstream" PC will believe that he or she is fully prepared for the
date changes in the next few years. Little do they know how far from the truth such
blind faith is.
THE FUNDAMENTALS
In order to address the Y2K problem as it relates to PCs, it is first necessary
to understand a few of the fundamentals of PC operation. We can group the various
levels of Y2K-sensitive PC features into 4 basic categories: the hardware (including
CMOS, BIOS, and system clock); the operating system (including both underlying O.S.
as well as add-on shells such as Windows); the applications; and the data (stored
data files as well as data interchange). A fifth level of Y2K sensitivity is the
user himself or herself. Users must learn to understand the long-term impact of
using two digits to represent a four-digit number. However, the user element is
beyond the scope of this article. This discussion will focus solely on commercially-sold
hardware and software products and how these interact with each other and with their
electronic environment.
Probably the first obstacle to be overcome in discussing Y2K is to rid the typical
PC user of the notion that Y2K compliance can be ascertained simply by observing
the PC's system clock. Most users who have some degree of awareness of Y2K issues
are woefully misinformed about this fact. They believe that if they push the clock
forward to December 31, 1999, and watch the timer roll over successfully, that this
is a sufficient condition to prove Y2K compliance. That is simply not so. The user
who mistakenly thinks otherwise must be told in clear and firm terms: "THE
APPLICATIONS DO NOT ASK THE SCREEN WHAT TIME IT IS." Applications ask the operating
system, which asks the hardware (BIOS, CMOS, or clock chip) what time it is. Occasional
exceptions to this sequence do exist, but the most important fact to explain to
the customer is that SOFTWARE DOES NOT ASK THE SCREEN WHAT TIME IT IS.
What every PC must do correctly is the following:
1. Hardware: generate, store, increment, and communicate the correct time (including
during power-off condition)
2. Operating System: read (from hardware as well as user input), store, increment,
and communicate the correct time (including power on/off transitions)
3. Applications: read (from the operating system and/or direct hardware read), store,
and communicate the correct time (some applications may also implement an increment
feature)
4. Data: read, store, and communicate the correct time.
Looking over this list, that makes at least 14 specific states that can affect the
Y2K compliance of a PC as a working system. Most vendors look only at the 4 states
in the Hardware item. A few users may be interested in verifying the 3 additional
states in the Applications item. Even then, they may only recognize the need to
verify one or two vital applications, ignoring the possibility of a system-wide
regression caused by a single non-compliant application. But even assuming that
all concerns on levels 1 and 3 are addressed, this still leaves HALF THE POTENTIAL
Y2K FLAWS unaccounted for.
Addressing the data issue on level 4 is tricky. Some applications have been "patched"
to shift the date data effectively 5, 10, 20, or 30 years into the past, from the
software's perspective. Some applications use one "windowing" size, other
applications use a different size. Some programs may even use an oddball window
size like 7 years or 22 years. There may be a need to interchange data with other
computer systems that have been date-skewed in order to implement windowing as a
temporary fix for an unrepairable Y2K issue. Which "windowing," exactly?
It varies from system to system. Imagine importing data from Lotus 123 and integrating
it with mainframe data from a 20-year-old system. Then e-mail the resultant data
to an accounting and tax office running Microsoft Excel. What assumptions about
date data are made by each application? Does everybody document this information
correctly and make it available and understandable to the endusers? Will the programmers
and managers who made the windowing decisions still be working for the same companies
one, two, or five years from now? Each application has its own set of windowing
assumptions and its own data format. How will endusers determine the difference
between date-sensitive corruption due to file format inconsistencies, versus misapplied
windowing ranges?
This is the type of situation that Microsoft will likely exploit to recommend an
all-Windows, all-Microsoft environment -- not just within each company, but across
the board between whole systems of suppliers, support companies, consultants, and
other clients. But what if the Microsoft products -- as they typically do -- cause
new problems that ruin the entire supply chain? Who will pay for all of this?
Can you begin to see why the computer industry and other high-tech corporations
are so desperate to push through laws giving them significant immunity from Y2K
lawsuits? Read the article HERE.
THE OPERATING SYSTEM AND Y2K
Notice that level 2 of the PC has been ignored until now. The operating system bears
the greatest burden of Y2K compliance. An O.S. must work properly with the system
hardware, even during challenging periods such as power-up and power-down. It must
correct or at least warn of incorrect date data. It must communicate such data to
a wide variety of PC applications, many of which were not designed specifically
for the particular O.S. in question. For example, a Windows98 user may need to operate
a DOS-based accounting package as well as a Windows 3.1-based spreadsheet. Or, a
tax office may need to use several different years of tax applications that were
designed for Windows 3.1, Windows95/98, or WindowsNT. Will the date data be exchanged
correctly with each application, both before and after the year 2000 rollover? Will
the O.S. know if an application attempts to "correct" the system date
erroneously?
Furthermore, operating system "patches" and "updates" are now
a standard feature of the PC environment. Who can be sure that a "free update"
or a "bug fix" will not cause a date-sensitive regression when it is applied
to the PC? What of the new generation of PC viruses -- will one or more of them
alter a date-sensitive feature of an operating system, causing a silent but deadly
regression to a non-Y2K-compliant condition? What about application-specific patches
and viruses; will a change occur which silently renders one mission-critical application
noncompliant, while the rest of the system seems to function normally?
Frankly, these issues are just too worrisome for most members of the PC media to
discuss. No doubt some of them remain silent out of fear of causing a "deer
in the headlights" response by the PC-buying public. Some writers and commentators
are admittedly weak on the technical side and cannot grasp these complex issues.
Possibly a handful of them are so thoroughly smitten with Microsoft that they do
not realize they are embracing death, just like many non-technical magazines eagerly
pursued the advertising dollars of the tobacco companies and enjoyed consuming their
sinister products.
RUMBLINGS OF DISTANT THUNDER: THE Y2K BUGS REVEALED
It is only in the last year or two that the grim, dirty realities of PC Y2K operating
system errors have begun to emerge. We can surmise the true condition about two
Microsoft operating systems based on the historical evidence:
1. MS-DOS. Probably non-compliant. Microsoft's DOS code is licensed to IBM, who
sells it as PC-DOS. IBM now admits that PC-DOS 7.0 (equivalent to Microsoft's latest
DOS) needed to be patched in order to achieve Y2K compliance. The IBM patch is available
HERE. IBM now sells
a Y2K-compliant DOS called PC-DOS
2000. However, Microsoft does not sell such a product. Microsoft will likely
not get serious about fully fixing MS-DOS, since DOS is too open of a development
platform for Microsoft to control with their leverage tactics. Microsoft would rather
use the fear of Y2K non-compliance to force uninformed people to downgrade unnecessarily
to closed, proprietary platforms such as Windows.
2. Windows 3.X. Probably non-compliant. The Windows 3.1 code is licensed to IBM
for inclusion in OS/2 Warp as "Win-OS/2". IBM did not finish patching
this code for Y2K compliance until the middle of 1998. These patches are currently
available as features of Warp 3 FixPack
40, and Warp 4 FixPack
10. Microsoft likely will not get serious about fully fixing Windows 3.0, 3.1,
or 3.11, since they are intent on reducing shipments of these products and selling
higher-margin operating systems such as Windows98 and WindowsNT.
The Y2K status conditions of other Microsoft operating systems have been occasionally
mentioned in the PC press recently. We can surmise the true condition about three
other Microsoft operating systems based on the articles quoted below, as well as
the historical trend:
3. Windows95. Probably non-compliant. Microsoft recently admitted the product needs
a patch. Read the article HERE.
However, several new Y2K bugs have appeared since then in Windows98 (which is closely
based on the Windows 95 code). Read the article
HERE. Microsoft
will likely not get serious about fully fixing Windows95, since they are intent
on reducing shipments of Windows95 and selling later operating systems. In fact,
MS Office2000 is reportedly a nightmare to install on a Windows95 system, indicating
that MS does not intend for Windows95 to continue as a viable O.S. (This behavior
was reported in a syndicated computer column featured in the Nashville Tennessean
of June 13, 1999.) Furthermore, compare the two date-range sizes used by Microsoft
as windowing patches in the articles HERE
and HERE.
How can Microsoft products reliably handle Y2K data interchange, if they are not
even consistent with each other?
4. Windows98. Probably non-compliant. Microsoft recently admitted that several Y2K
bugs are embedded in Windows98 version 1.0. In fact, one article quotes a French
government minister as considering prosecuting Microsoft for consumer fraud; one
of the Windows98 installation screens clearly claims that Windows98 is "ready
for Year 2000" even though this is blatantly wrong. Read the article HERE.
Now Microsoft is delivering "Windows98 SE" which is really Windows98 version
2.0. Read the article HERE.
Version 2.0 is supposed to include Y2K fixes, but Microsoft is not advertising that
fact strongly because it does not want to point attention to the fact that perhaps
50 million copies of Windows98 version 1.0 are currently in circulation and in use
-- and are non-compliant for Year 2000. Is W98 version 2 really ready? Given Microsoft's
constant wavering and waffling, as well as regular discoveries of new Y2K bugs,
the only safe conclusion is to say "probably non-compliant."
5. WindowsNT. Probably non-compliant. One unusual feature of WindowsNT is that it
attempts to wrest control of the PC's date, time, and I/O functions away from the
BIOS at boot time and implement them in software. This is one reason why WindowsNT
is notoriously unreliable, particularly on PC hardware that has not specifically
been verified by Microsoft as acceptable to WindowsNT's fragile infrastructure.
Microsoft had long insisted that Windows NT 4.0 was Y2K-compliant, but we now know
this was not so. A patch called Service Pak #3 (SP3) was claimed by Microsoft to
solve these problems. However, Microsoft was embarassed to admit in early 1999 that
this was in error. Read the article HERE.
SP4 theoretically fixed WindowsNT 4.0, but that also turned out to be wrong also.
So it is now mid-1999, and WindowsNT 4.0 needs SP5 in order to meet the current
claims of Y2K compliance. Read the article HERE.
Is NT4 with SP5 really ready? Given Microsoft's poor track record and new Y2K bugs
being discovered every few months, the only safe conclusion is to say "probably
non-compliant."
There are several reasons to believe that Microsoft will continue to find new date-sensitive
bugs in its products. For example, *every version* of Windows fails to correctly
implement the Daylight Savings Time algorithm. Read the article HERE.)
It is apparent that Microsoft's testing and regression-prevention methods are abominable.
It is very likely that future "fixes" will cause regressions (slipping
backwards) to occur, making new Y2K bugs appear out of nowhere on systems that may
have previously been compliant or nearly compliant.
Of current operating systems, this analysis leaves only Windows CE remaining to
examine for Y2K compliance. I have not been able to find a single article that discusses
whether or not WinCE is compliant. Nobody seems to have even questioned it, despite
Microsoft's miserable record of errors, bugs, defects, and flaws leading to a butchered
record of Y2K noncompliance for desktop operating systems. Furthermore, CE is a
ROM-based O.S., meaning that a software patch may not be sufficient to fix any Y2K
flaws that surface.
Microsoft is supposedly discontinuing WindowsNT and delivering a different operating
system called "Windows2000" sometime in the next few years. Once again,
I have not been able to find a single article anywhere that discusses whether or
not Windows2000 might be ready for Y2K. The PC media seems to have developed an
asteroid-sized blind spot with regards to so-called "new" products and
Y2K. Millions of lines of old, legacy WindowsNT code are being mixed with tens of
millions of lines of untested code to form this Windows2000 product. Read the article
HERE. There is no long-term
track record for Y2K compliance, because the product is not even available in finished
form. It will not be available in time for a thorough, long-term Y2K shakedown.
It is a risk that is not worth taking.
ENSURING A Y2K-READY PC ENVIRONMENT
The first step to Y2K preparedness in personal computing is to verify the hardware
compliance. The IBM test application is ideal for this verification: it is free,
it is thorough, and it clearly explains the results and makes recommendations. Yet
the entire test procedure takes less than two minutes per PC. The IBM Y2K BIOS tester
is available HERE.
The second step to Y2K preparedness in personal computing is to completely back
up all operating systems, software applications, and data that are necessary to
continue operating the PC. The purpose of the backup process is to ensure that in
a worst-case scenario, the PC clock may be set backwards and a complete system restore
accomplished in order to resume operations in a temporary mode until remediation
is accomplished.
The third step to Y2K preparedness in personal computing is to remove all Microsoft
products from every computer in the entire business enterprise, as well as any home
computers and laptop/notebook computers used to handle date-sensitive data. Date-sensitive
data may include Internet access, particularly e-mail features, due to timestamping
required by Internet protocols. All other non-compliant software applications should
be removed as well. If there is some doubt about the compliance of a particular
application, the application's manufacturer should be consulted. The only personal
computers which do not need Y2K testing are:
1. PCs used solely for games in a non-networked, non-Internet configuration;
2. PCs used solely for standalone word processing or graphics design with no need
for Internet access.
"Standalone" means non-networked computers; any computer network may propagate
a bad date timestamp or erroneously-dated application data to another machine on
the network, which may cause a Y2K failure elsewhere on the network. It is therefore
VITAL that ALL NETWORKED PCs be similarly configured and verified for Y2K compliance.
The fourth step to Y2K preparedness in personal computing involves building a new,
Y2K-compliant information infrastructure. Once a computer has been cleared of non-compliant
applications and operating systems, an alternative configuration using 100% Y2K-compliant
operating systems and applications is required. In large installations where downtime
is not permissible, it will be necessary to build up the alternative system in PARALLEL
with the old, non-compliant system, and then schedule a transition period. Now more
than ever, "hardware is cheap." Build a second computer or a second network,
isolated from the system that has questionable compliance, and copy the data over
first. Verify full compliance of the new system, and then perform the operational
transition.
The fifth and final step to Y2K compliance is to Test, Test, Test the finished systems.
RECOMMENDED OPERATING SYSTEMS AND APPS
For Y2K compliance in a general-purpose PC environment, I can recommend only two
commonly-available operating systems at this time: either Linux, or IBM OS/2
Warp. Linux is not known to have any Y2K-related issues, but it is scheduled
to have a Year 2038 problem. For most PC environments, this should be enough lead
time to allow for a fix to be developed. However, that assumption was what got the
whole computer industry in trouble many years ago. Some people may not want to transition
to a platform and build an enterprise now, only to suffer the same indignities and
dangers many years from now. IBM OS/2 Warp 4.0 with FixPack 10 (FP10)
meets the IBM standard of successful operation through the year 2099. Warp 3 can
be made similarly Y2K-ready with FixPack 40 ( FP40).
It is quite possible that OS/2 will work successfully for another 900 years, but
IBM has not specifically guaranteed that at this point.
However, the main reasons to favor OS/2 Warp over Linux at this time are as follows:
1. OS/2 Warp is easier to install and blend into most standard desktop environments;
2. OS/2 Warp has a single codebase maintained and regularly updated by IBM;
3. OS/2 Warp runs a wide variety of DOS, Windows 3.1, native OS/2, and Java applications;
4. OS/2 Warp has a consistent user interface that is very easy for Windows users
to learn (usually one or two days);
5. OS/2 Warp has a well-established reputation for long-term reliability in a variety
of environments.
In the future, it is possible that Linux may surpass OS/2 in one or more of these
areas. However, at present OS/2 Warp is in my opinion the better choice for most
home and small business PC users.
Several OS/2 applications suites are available including Lotus
Smartsuite 1.1, Star Office 5.0,
and a family of desktop applications from Sundial
Systems. As far as I am aware, these applications are fully Y2K-compliant. Thousands
of other OS/2 applications are listed at the SoftWhere
website. IBM has also compiled a list of Web links to information about certain
third-party applications and their Y2K compliance levels. Read the article HERE.
CONCLUSION
For some people, the whole idea of Year 2000 readiness is a waste of time. Most
gaming software and many educational, word processing, and graphics applications
have no date sensitivities to worry about. In many home PC environments, Y2K readiness
is simply not an issue. However, anyone who handles PC data in a business environment,
or who uses Internet e-mail, or who does any kind of accounting or scheduling work,
MUST consider the real threat to their business and personal financial condition
if they do not address the Y2K compliance issue NOW, while there is still time to
act. Also, we must keep in mind that January 1, 2000, is not the *end* of the Y2K
problems, but rather the beginning of a new level of errors and crashes. One major
study predicts the Y2K problems to persist for a period of up to 30 months. Read
the article HERE.)
Therefore, there are real long-term benefits -- not just a singular point of safety
-- in recommending and upgrading yourself, your friends, your co-workers, and your
clients off of the doomed Microsoft platforms and up to IBM OS/2
Warp.
While Microsoft's website finally has some information about the Y2K defects in
their products (after several years of stonewalling and denial), I find it hard
to accept that these statements would be accurate or truthful. One reason to doubt
the veracity of Microsoft's "party line" is that it has changed so often
in just the past few months. Another reason to question Microsoft's honesty is that
their statements during the recent Department of Justice trial have been clever
and slick, but not always accurate. Even their own carefully-crafted video demonstrations
have been exposed as rigged and therefore untrue.
Meanwhile, those of the PC press who have lived in a dreamworld for the past ten
years, mindlessly following Microsoft with cult-like blind faith, must awaken from
their intoxicated stupor and face reality. Microsoft products have real, proven,
and admitted Y2K defects. Millions of PC users are in darkness about these problems.
None of the "mainstream" PC writers has had the guts to boldly warn people
away from Microsoft products and upgrade them to Y2K-compliant environments such
as Linux and IBM OS/2 Warp. As
a result, the "mainstream" PC press is due for a grave downfall from their
phony pedestal. Those of us who are wise enough to recognize the impending demise
of the "mainstream" computer press should prepare now to produce viable
alternative publications after the Millennium rollover.
As for Microsoft itself, it is quite possible that the whole DOJ lawsuit may soon
be moot. Less than one year from now, Y2K anger may possibly cause the total collapse
of the monopoly. What of Microsoft's constant backtracking about Y2K, its never-ending
discovery of new bugs and flaws in its products? Is Microsoft intentionally lying
when it claims Y2K compliance for some of its products, or is Microsoft so totally
incompetent that it does not even KNOW whether its products are compliant or not?
Does it matter?
