Did anything happen on April 1, or April 9, that shut down any systems anywhere? April 1 was suppose to be a significant date for some failures, and April 9 was the 99 day of the year and could have caused problems. If nothing happened and nothings hapens on sept 9 99 (9/9/99). Then are we to assume that when the year 2000 hits not much will happen either.

-- Dennis Tolliver (tolly@dtgnet.com), April 25, 1999

Answers

4/9/99 and 9/9/99 have been way overhyped. They are only problems when programmers explicitly decide to use them as end dates. That was old-fashioned even 20 years ago the last time I saw anyone do it. Especially when you could have used 99/99/99 or 99999 (julian) and put in checking for it. Those are not valid dates at all (although they will still cause sorting problems when 00/01/01 or 00001 roll around!)

As for April 1, the speculation was that when a fiscal year starts, a program would possibly need to use the end-date of the fiscal year (April 1, 2000) for something, and crash. This actually did hit state unemployment agency software back on Jan 1. But these fiscal year calculations are a tiny fraction of all date/money handling software, and have nothing to do with embedded systems.

As for serious failures, expect any organization to run into trouble when they start taking orders for dates in 2000, or sending out orders scheduled to be filled in 2000. Or after Jan 1, when their systems try to bridge the 1/1/00 date. And most embedded systems will crash on the 1st, or soon after.

Sorry, but there's not going to be a big crash all at once that people can see. You are either going to have to educate yourself about this stuff and trust your own judgement, or pick an expert and go with them.

If you want background info, see my article, off my home page at http://www.best.com/~mgoodfel

-- Michael Goodfellow (mgoodfel@best.com), April 25, 1999.

Good morning Dennis,

I agree with Michael that the 4/9 and 9/9 problem was overhyped. I did a survey here a while ago, and we have dozens of programmers here, with more than 875 man-years of programming experience. Only one (I believe it was Chuck?) mentioned that he knew of systems where this could be a potential problem.

As for other failures in 1999, I'll give you my rubber stamp response. They call it the Y2K problem for a good reason, not the various dates in 1999 problem. The number of programs that do look ahead processing is tiny, when compared to the total number of programs that have a date problem. Those that do are well known, and would be fixed first because they are needed first. And I believe that most embedded systems failures will not occur until 2000. <:)=

-- Sysman (y2kboard@yahoo.com), April 26, 1999.