The Y2K bug wasn't created by programmers' errors

Tuesday, February 01, 2000

A genuine problem

The Y2K bug wasn't created by programmers' errors


   THE ROANOKE Times' coverage of issues related to the Y2K rollover bug in computing systems has been unusually irresponsible, even by current journalistic standards, the essence of which seems to be that political correctness is more important than knowledge and that a sensational headline is more valuable than accuracy.

    As one who worked several six- and sometimes seven-day weeks on this problem, I am disappointed that you have chosen to treat the fact that disaster did not happen as if it were a demonstration that the whole thing was one big false alarm. The day you published your front-page headline about Y2K having had no sting, a colleague at Information Systems and Services was told by a client that that the whole thing was a hoax -- after we had spent several long days saving this client from ruin.

    Then, on the Jan. 9 Commentary page, you published an opinion piece ("Thanks, Y2K fixers -- but just who bred the bug?") by an ignoramus who said, in effect, that the problem was the fault of the same people who were fixing it.

    I will admit that people who are too lazy to do any research, or too stupid to ask someone who has knowledge of the subject they are writing about, still have a right to publish anyway. The columnist, Denis Horgan of The Hartford (Conn.) Courant, clearly wanted to do a little geek-bashing, so any explanation made to him would be a waste of time.

    He didn't ask me what I thought about Y2K, but in the last few months a lot of other people willing to consider the matter thoughtfully have asked. The substance of their questions covered two of the points made in the article: How serious is the problem? And, how could you guys have made such a mistake? Here is a condensed version of my answer:

    (1) It is very serious.

    There is broad agreement among knowledgeable people that Y2K is a threat. There are some differences on specifics, which can be accounted for by the varieties of outlook inherent in human nature and by the differences in areas of expertise. But I do not know a single expert in the field who thinks Y2K is anything less than a problem that must be fixed to avoid serious harm.

    (2) It wasn't a mistake

    There have been a lot of changes in the last 30 years in information systems that no one could have predicted. One of the most important, and least understood by the lay public, is the effect of cheap memory and storage.

    If the subject of the century rollover had been considered in the 1970s, we would have done the same thing anyway. The reason: Memory was so expensive that any optimization was extremely valuable.

    When I started, it was common for programmers to spend days working for a few bytes reduction of memory requirement. Most people in the field today have no clue about the difference it makes to have memory measured in bytes per dollar instead of dollars per byte.

    The latter situation existed in the early part of my career. Multiplied over the large number of machines and programs in which it existed, using two-digit dates saved millions of dollars.

    The real problem is that the practice continued long after memory became cheap. The original reason for two-digit dates was gone, but they continued to be used for the same reason many traditions exist: Change does not occur without demand for change.

    Perhaps Horgan would be so kind as to write an explanation of why he didn't publish an article years ago pointing out that this change needed to be made. I'm not going to hold my breath waiting for it.

    The new-century date problem may be new to Horgan. It isn't new to people in this field. Applications written here at ISS by Jerry Schenkel more than 10 years ago made provision for this event. The change was made here at an appropriate time, after memory and storage had become cheap enough to make it feasible.

    There is a well-funded, 40-year-old company here in Roanoke that did nothing to prepare for Y2K. We learned of this at ISS when its representative showed up at our offices with an urgent interest in retaining our services on the morning of Jan. 3. Anyone who does not believe the threat was real and substantial would get an education from talking to the folks there.

    Horgan mentioned an assisted-care facility that "redirected nearly three-quarters of a million of its precious dollars to replacing special locks, lest the fragile residents risk being caught behind paralyzed doors; prudent, particularly as there was no indication that the computer-chipped locks wouldn't work otherwise."

    I have looked into this matter a bit, and I too find no indication that locking devices with electronics integrated into them would quit working due to the date rollover. This was a waste of money. One wonders whether Y2K was used as an excuse instead of a reason.

    Perhaps Horgan should ask them why they did it. He can't blame the computer geeks for that one; it takes top management to screw things up that badly. We would have known better.

    It seems to me that this could be an example of a facility that employs the same sort of militant ignorance in its management that Horgan employs in his writing. It seems churlish of him to criticize what are, with near certainty, his brothers in spirit.

    HOWARD DeLONG is a networking analyst for Information Systems and Services in Roanoke.

Last updated 2000/02/02
Archived at the Department of Computer Science web site with the kind permission of:

in support of college education.