dargopolis

Re-reading my original comment half a day later, I realize that it is wildly inaccurate. I do remember the Y2K mainframe changes part of the project being kind of chill. However, that is only in comparison to the rest of the project. We were redeveloping all the other mainframe systems (and building a few new ones) for a completely different platform and it had to be done ASAP because Y2K! Though it was ultimately successful, it was a classic death march. Good times.

In the last half of the 90s, I was working on a project to move (redevelop) a bunch of mainframe applications to a client-server architecture. It was known that there would not be time to redevelop some of the required applications in time for Y2K, so they had to be updated to support 4 digit years instead of just 2. I don't recall any real panic as the work was started years in advance and probably completed with several months to spare. I had previously worked at that location in 1990 and by that time they already had a standard that any new development was to use 4 digit years (it was part of the code review checklist). It was only the oldest, pre-1990, mainframe systems that needed the fix before Y2K.

Having read Jonathan Schaeffer’s great book One Jump Ahead, I couldn’t help feeling that this article was a (well-written) book report which just barely mentions its primary source. Nice summary, but read the original. I haven’t read the second edition which apparently includes the follow-up story of solving the game, but I found the first edition very compelling. I wonder if the reporter was following the recent Deepmind/U of Alberta news and then came across the checkers story. Disclaimer: I was in Jonathan’s compiler course 30 years ago, prior to Chinook.