Year 2000 Concerns




Most laymen think of the Y2K problem as a simple hardware problem Hardware is however only a small part of the problem. The Y2K problem is actually very pervasive and can also be very subtle.


Hardware


For a PC to operate correctly in the year 2000 and beyond, its Basic Input/Output System (BIOS) must support the century change and not assume that a 00 in the date means 1900. The real-time clock device In all **86 PCs, the real-time clock counts only from 00 to 99 and then wraps around back to 00 without taking any further action. The century is stored in a non-volatile CMOS RAM cell which must be maintained by user and/or software. In addition to this, motherboard manufacturers sometimes modify a BIOS to make it work with their products. As 2000 approaches, most (note the word most) BIOS problems are being solved (in new machines). A recent test by Greenwich Mean Time (GMT) Corp. on 500 machines made in the second half of 1997, encountered a failure rate (Y2K problem) in 21%. This is down from 97% a year earlier but still rather high. However it may be noted that the type of failure is not noted and some may simply be a date problem that disappears on rebooting. There are diagnostic tools to determine if your system will have a problem. (You may also simply call your manufacturer.)


Software


The more difficult problem to solve is that associated with software: programs and data. Some programs are just not Y2K compliant at all. And while others are advertised as being totally compliant, there are different Techniques for resolving the problem and no standard for everyone to follow.

One interesting case is Microsoft Office. When GMT Corp. tested it, it found that the versions of Excel and Access in the package both used “windowing” to become Y2K compliant. However, they used different pivot dates. Therefore, if a user created data (including dates) in one program and ported it to the other, there was a possibility that it would be corrupted. A second example is Microsoft’s SQL server database that was made compliant by choosing a pivot date of 1950. A VA hospital using it found by chance that they had patients that hadn’t been born yet.

These are just 2 examples that have beef found before 1999. What will happen after Dec. 31, 1999 is anyone’s guess. It is common for users to do important data storage, analysis and manipulation on their PC rather than on a mainframe. It is simply much faster that way. This software problem though doesn’t only refer to commercial programs. What about all the “in-house” programs, macros, spreadsheets, etc. that abound. They must all be considered suspect.

Data corruption will probably be the most significant result of the Y2K problem. To badly quote an expert - by using programs such as VBA, OLE, and relational databases, you will be able to make the corruption occur and spread very fast in multiple dimensions. This situation is not a great deal different from a virus. One program corrupts a small portion of data which is placed back in the overall data base. At a later time it is used by someone else who either corrupts a different part or re-corrupts the corrupted data or just uses the incorrect data to analyze and create new (incorrect) data. This is a standard exponential growth and it is just this thought that is driving the Y2K frenzy today.


Side Issues


In addition to the above rather obvious issues, there a few other ones that will affect some areas quite dramatically. These are: Leap Year; 1999 Exception Handling; and Embedded micro processors.

Compared to Y2K, the Leap Year problem is, at best a hiccup. As a rule of thumb, years divisible by 100 are not leap years; 2000 however is. This small problem was also not put into the machines. Unless the 28 year setback technique for correcting the Y2K problem is implemented for employees records, there will be some problem associated by this extra day in the year. (ex How many hours did you work. How much should you be paid. Did you accumulate vacation time on Feb. 29. etc.)

The 1999 Exception will be here very rapidly. Many software programs have been using a series of 9`s in the date as an exception code. An similar series in any portion of the date field can be interpreted as an error. One example was a Drug dispensing machine in a Mississippi hospital (Jackson Mississippi Baptist Health System) that refuse to refill a drug prescription because it had a refill date in 1999. A related problem is the use of similar codes in the ISO-Latin-1 character set. One possible problem area is in the decimal code “199” which is used for capital C cedilla. The second code that could cause a problem is one of the undefined characters; decimal 153. In hex format, this is “%99”. This transmitted as “2599” (“25” is the hex code for “%”). Programmers have been in the habit of using the undefined character set to transmit special items such as EOF and EOR. A potential for ambiguity can easily be seen.

Embedded Micro Processors also have the ability to cause havoc. 8080s and Z80s are ubiquitous (and invisible). Devices all around us use them for control purposes. There is a story of a person that bought a new car recently and reset the date on the processor that controls it. The car “thought” more than 100 years had passed since the emission system had been tested and shut down. Offices and homes use chips every day. They are in postage meters, phones, environment control system, fax machines, security systems, microwaves, stoves, etc. If nothing else, it will be an interesting and possibly frustrating few years.


Related Problem
(Dow Jones Industrial Average)


Editors note: In light of this months stock market, this now seems like a lot less likely to be a problem.

Charles Dow, the first editor of the Wall Street Journal, and one of the founders of Dow Jones Company, is credited with first creating the averages which today are the most widely followed in the world. Dow originated the averages in 1884 as an attempt to express the general level and trend of the stock market. To do this, he chose what was in his perception, to be stocks representative of the general market. The Dow Industrial Index is merely an average of the prices of the 30 stocks of which it is comprised.

When the Dow Industrial Average was originally created, it had a base of 100. That is, on the day it was created, the Dow Industrial Average was 100. In 1966 it finally hit 1000 and in late 1982 it finally broke the 1000 barrier. Since then it has grown almost steadily. Between 1983 and the end of 1995, the Dow Industrials rose more than 3500 points. In mid July, 1998, the average was about 9,300. (Today 8/13/98 it is down to 8,522.84 but seem to be bouncing back up).

Many financial programs were designed with the assumption that the Dow Industrial average will never rise about 9,999. When the programs were written, that seemed like a reasonable assumption. It looks now like this magic number will be reached sometime in the year 2000.