It was just one of twenty panels on a graphic celebrating the 20th anniversary of TechEd in Australia. But it made my blood boil. “’00: Y2K does not happen. Life goes on.” I expect that kind of uninformed trolling on forums. I don’t expect it at a technical conference.
If you’re too young to remember, the Y2K problem, often referred to as the millennium bug, was the challenge of dealing with older software which wasn’t designed to recognise dates after 2000. Space was absolutely at a premium in the first few decades of computing, so coding to recognise four-digit dates simply didn’t happen. That didn’t seem like a problem in 1970 or earlier, but it was a problem once 2000 hit, since you could no longer assume the first two digits of a date based on the last two digits. The behaviour of some of those old systems was unpredictable, and in many cases the COBOL coders who had worked on it had long since retired.
I began working as an IT journalist in 1994, and how to deal with the Y2K bug was one of the most persistent topics we covered. Changing those entrenched systems was fiddly, expensive and time consuming, but the process started early. As a result, by the time 2000 actually rolled around, most major issues had been dealt with.
Prior to Y2K, the most dire predictions suggested that most utilities in the world would fail and we’d all be forced into a survivalist scenario, eating baked beans and plucking spare clothing from corpses. Post-2000, a different tinfoil hat scenario emerged: that the entire thing was a con, designed to give cushy work to programmers, and that if we’d ignored them, it would have made no difference. That’s essentially the attitude conveyed in the image pictured.
It’s stupid, it’s dishonest and it’s wrong. There wasn’t a major issue with the Y2K bug, but that wasn’t because the potential didn’t exist. It was because numerous IT workers put in significant effort to make sure their software would function correctly. That doesn’t deserve mocking or a casually dismissive attitude.
Date-related problems frequently occur in computing. For instance, many 32-bit systems may experience problems when we reach 2038. In Excel, you can’t perform calculations on dates earlier than 1 January 1900. The complex and arbitrary nature of dates (uneven month and year lengths plus calendar changes) mean this will probably always be the case. We have to code in a way that’s mindful of these issues — and we shouldn’t pretend that never happens.
Disclosure: Angus Kidman travelled to TechEd Australia as a guest of Microsoft.