[CSUSB] >> [CompSci] >> [Dick Botting] >> [Papers & Essays] >> rjb99d.y2kbau
[Index] || [Contents]
Disclaimer. CSUSB and the CS Dept have no responsibility for it.
This is a Beta Version, read at your own risk.
Copyright. Richard J. Botting(Sun Jul 6 10:03:44 PDT 2003). This paper is being developed for publication. Permission is granted to quote parts of it as long as the source is acknowledged and the author informed.

Contents (index)


    Y2K is Business as Usual

    This morning I started my laptop. It displayed the desktop. Then it crashed and rebooted. This happens on about 20% of all the time. I do not have to wait for the end of the millennium! It crashes by the end of most weeks.

    I supervise a network of UNIX boxes. We have about 300 people who use this network. UNIX will have a problem in 2038. I do not have to wait for it. When I dialed into my CPU, I found that it could no longer access email. The new mail server no longer supports Networked File System file locking. Half the UNIX systems that we have use this feature. We had been testing the new system for some weeks. We discovered the missing feature almost immediately it replaced the old server. The old server had a mean time to fail of about ten days because of bugs in its name daemon systems. Malformed packets, bad interrupt handling, counters overflowing, zombies filling process tables, incompatibilities between releases and versions, and disks filling up with erroneous error reports add up to a mean time to failure of about three weeks. I do not have to wait for the end of the millennium for these systems to crash. I get nervous when one runs for more than 100 days without a problem.

    When I explained the latest network problem to my wife who also uses the same Email server -- she is a cybernetician, a Macintosh user, and a Beta tester for my network -- she remarked: "Who needs to worry about Y2k? Things are already buggy." It took about thirty minutes to change her software so she can read her email.

    In the evening, my laptop froze within three minutes of starting to edit this note. The touch pad and keyboard were dead. The only button that works is the power key. I think there is a bug in the touch pad driver. It happens when my finger crosses an invisible line at the same time as a button is clicked. This happens once a week or so. I think it comes from failing to simulate a mouse wheel on a flat pad. I hate it. It smells of the kind of thoughtless feature that I have (during 40-years of programming) often patched into programs. Other defects stop me from disabling it. (1) The control panel claims to have turned off the feature but it is still on. (2) The supplier of the operating system has the wrong phone number for downloading updates. (3) The modem+local phone lines has a 30 minute mean time to failure on binary downloads. I put up with weekly crashes.

    Shortly after the above touch-pad crash, my wife's spell checker program started to misplace corrections and misinterpret wordprocessor codes. Avoiding, allowing for, and working round bugs is normal for us. It is typical of our life with computers. Manifestations of artificial stupidity; bugs, crashes, and misfeatures; are countered by our intelligence and creativity. I expect most other computer users use their real intelligence to find ways to work around artificial stupidities. The Y2K problem is another. Working round it will be business as usual.

    My wife, added another prediction to the above. She recalls that when countries shift the side of the road they drive on, the change occurs with no accidents. The accidents happen a month or two later, when the real intelligences forget the change. She predicts that 2000 will start with no more than irritations. The disasters will be embarrassing and occur in February 2000.

    Does this mean that computer professionals can ignore the Y2K problem? No, Y2K is calling for heroic efforts. But is this unusual? Not at all! In World War II the GIs had a significant acronym: SNAFU. This means (approximately) "Situation Normal. All Fouled Up." The existence of heroic "death march" projects is well documented. Most software processes are immature. Immature software development always involves last minute fixes and twenty-four hour work days. Time pressure leads to mistakes. These may get fixed. More often the users find a way round the problems. They get fixed in the next release - may be. The Y2K SNAFU is just Business As Usual.


Formulae and Definitions in Alphabetical Order