My recent columns on the
Year 2000 problem, in which I suggested we
just might have a real problem, drew a
lot of comment — much of it, as always, a
lot more interesting than my own.
From a friend at the FDIC (the Federal Deposit
Insurance Corporation):
Did you notice the recent announcement from the
Fed that it will increase cash reserves at its
banks in '99 in anticipation of greater demand
for greenback withdrawals late in the year?
I'd say that is a major wake up call to folks to
do a little common sense preparation today.
I'm not a market timer, but I've already (back in
May) moved the bulk of my retirement portfolios
out of equities and into bonds and governments.
I plan to dive back in to the stock market sometime
in 2000 after either the dust clears, or it appears
there is much to do about nothing. I figure at
worst I'm trading potentially greater stock run-up
for lower bond yields for 18 months, and in my
"best case" I've conserved my principal
and accrued gains of the last few years and can
buy back into the market when stocks looks cheap
again in 2000.
By the way, credit card servicers are busting their
guts to get ready for Y2K. Many of them are not
accepting, or delaying for 6 to 8 months, conversion
accounts from other servicers. I've learned this
from a recent bank failure that involves hundreds
of millions of dollars in credit card portfolios.
(The WSJ is working up an article about this failure
as we speak.)
From Emmett Redd, Associate Professor in the
Technology Department at Southwest Missouri State
University (I had said that if the power went out,
so might the ability to pump gas):
Gasoline probably won't be too hard to pump without
electricity. On many pumps I am familiar with, just
take off the side, loosen the belt between the
electric motor and spin the pump by hand. Also,
most of the tanks around here are above ground (to
avoid groundwater contamination). Just shut the
valve, "break" open the lines and gravity
fill whatever container you have available. The
store owner will like your bag of dimes better than
the muzzle of your shotgun. :)
However, I have lived through enough ice storms where
electricity has been out for 2 or 3 days for neighbors
of mine. Even our house got cold after 16 hours without
the forced air furnace. Therefore, I have always lived
my adult life with at least two sources of heat, at
least one of which does not require electricity. Wood
has been a great primary source for over half of my life.
But now my house has a propane heating stove (with a
300 gallon tank out back), baseboard electric (in case
I forget to call the propane truck), and a kerosene
heater (purchased back during the high fuel price days
of 1985, but now used to keep the apple trees frost-free
or to warm up the diesel tractor for starting when
electricity is not available).
The advice for multiple heat sources applies for
every winter here. And don't forget
Jan. 1, 2000 is likely to be cold.
Most of the rest of your list is fine, but
Jan. 1, 2000 is too close to the solstice
for a solar water heater. Just heat the water on a
non-electricity dependent heater.
Y2K or not, the solar electric fence charger should
(and is) keeping my cows in the pasture they are
supposed to be in.
Be Prepared.
From Kevin Rasmussen:
A pretty in-depth exploration of personal Y2K
preparations can be found at:
http://millennia-bcs.com/prep.htm
This is a part of the Cassandra project web site,
which is a non-profit organization that discusses
"Y2K and the Risks to Public Health and Safety."
It's a neat source of information and anecdotes.
From John Wildenthal:
There are a few verified examples of embedded chip
failures. You might look at
www.cv.nrao.edu/y2k/sighting.htm
for some verified problems, including the valve at a
UK nuclear plant (scram/failsafe in 20 seconds after Y2K).
Power generation (except for windmills, hydroelectric,
and solar) is a process of boiling water-spinning
turbines-cooling water, over and over. The fan
problem you discussed is evidently not uncommon in
automated water valves. DeJager's Damocles site
mentioned that in the context of a sewer plant until
DeJager shut Damocles down.
A.T.:
I'm not entirely certain what this means or who
DeJager is. Damocles I knew in a previous life.
(High school, that is.)
From Jonathan Hochman:
No doubt there will be some Y2K disruptions. Last
year we had an El Nino weather pattern. How many
times did we hear about various problems that were
caused by El Nino? Lots of people started joking
that every little problem must be the fault of
El Nino. Those inclined to 'pass the buck' will
take advantage of the year 2000 to blame all sorts
of failures and lapses on mysterious computer bugs.
Having spent a considerable number of years earning
a computer science degree, I can give a quick lecture.
Binary computers (virtually every device that people
commonly call a computer) store numbers as 0's and 1's.
Depending on how many switches (bits) are allocated to
store a number, the number can range from 0-1 (1 bit),
0-3 (2 bits), 0-7 (3 bits), 0-15 (4 bits), and so on.
Some legacy programs, written in archaic languages
like Cobol and Fortran, may show two digit years when
they print or input dates. However, I see no reason
why the programmers would have wasted their effort to
program special arithmetic functions valid from only
0-99. In 1999, a year stored as an 8-bit number (0-255)
will be stored in the computer as the binary equivalent
of 99. When the computer adds one to the year, the
computer's 8-bit storage slot will contain the binary
equivalent of 100. Maybe the computer will print '00',
but in memory the date will still be valid for many
arithmetic functions like adding, subtracting, and
comparison. The program's math and logic functions
might continue to provide correct results for up to
155 years!
Recently the FAA determined that its non-Y2K-compliant
air traffic control computers would fail in 2007 because
they stored dates as (I think) 5-bit numbers, 0-31.
Maybe the software was implemented in 1975, so 1975 was
represented as zero. Failure is expected in 1975+32=2007.
I doubt that many essential devices care about the year.
Clocks in computers count up from 0 to 2*2*2*2*2*2*...*2-1.
The year 2000 is nothing special to them, because they
use binary math, not decimal. They could 'flip' (like a
car odometer) any time, depending on when they started
counting, and how many binary digits they store.
Logically, we should expect old computer software and
chips to produce unexpected date-dependent results from
time to time, whenever the date exceeds its storage
range. We probably have been experiencing these
problems all along. In 2000, the biggest problem will
be that various systems will print out funny dates, or
fail to understand newly input information. However,
it does not follow that math based on dates will fail
to calculate properly. I very much doubt that the power
grid will go down, or that digital toasters will suddenly
start burning toast because they no longer know what
time it is!
Please put the problem in perspective, and do not be
deceived by those who stand to profit from high
consulting fees to re-engineer old software. Anybody
using such old software should probably replace it
with a modern product. I recently bought Oracle
stock (ORCL), because Oracle sells complete software
packages for running many different kinds of businesses.
Why fix the muffler on an old junker when something
else is going to break next week? At some point its [sic]
better to replace the whole car than to keep spending
money on incremental repairs. Anybody in big business
worried about Y2K can call Oracle and buy a sparkling
new information system, free of Y2K bugs, much less
costly to maintain, and providing all the latest
features to enhance efficiency. (For small business,
I suggest Peachtree Accounting, which is also Y2K
compliant.)
A.T.:
I hope Jonathan is right, as he well may be. It's
certainly true that a great way to solve many Y2K
problems is to get a new system altogether. I'm told
this is why many big companies have gone or are thinking
of going to an SAP software system, a hugely elaborate
but elegant accounting/inventory/everything system
that works wonders. One small problem: The system
takes about two years to install. Let's see. What's
August 26, 1998 plus 24 months ....
From Jerome Payne:
I read your article on Y2K today and I find it
amazing that people are comparing this to a hurricane
or a blizzard or an earthquake. When is the last time
any of these events impacted the whole world at the
same time (relatively speaking — 24 hours to be
precise). There is a major difference between Y2K and
any natural disaster we've faced to date — there
won't be the rest of the country/world to help dig us
out of it, they'll be digging themselves out.
Hope you will rethink the magnitude of this occurrence.
The closest natural disaster this will be like would
be for an asteroid 1-mile in diameter or so to impact
the earth. The whole world would feel its impact
— severely.
A.T.:
I highly doubt it will be terrible, which is why I
suggested there was an 80% chance 2000 would come
and go with little more drama than Kohoutek's comet.
But it would be very foolish not to take simple,
cost-effective precautions just in case. (Buying in
bulk is a good idea anyway. Having an alternative
energy source can be a good idea anyway, although
it's obviously not practical for everyone.)
More of your comments to follow.
PS - Thanks to Peter Yeates for pointing out that
"Warren Buffet has sold his zero coupon bonds
— see
www.berkshirehathaway.com/qtrly/2ndqtr98.html
Note 4." What a nice little trade that was.
Why didn't I put $4 billion into long-term
zeroes when long-term rates were in the 7% range?
But if he's out now, that may tell you something,
too.
