Yes, sure.

Anyway, this time it will be different.

That time, people were aware of the possible problems - certainly more than it was necessary, but they were and that's the point.

This time, people will generally be unaware of any potential problems, and it will not be a midnight between two basically free days...

And there are more such "rollover" problems around. Actually, they were around years before Y2K. I came accross it for the first time I did no actual programming yet - as a secondary school student we were having a "practicum" at a local company's mainframe computer, and they gave as a long long listing of assets to correct apparent errors in the data - the most common was that the current value of some buildings was BIGGER than the original value, as they used only the last 2 digits of the year of build - e.g. 97 (this was in 1987) in some cases meant 1897, or 1797, and in one case even 1697...

One thing is true, people are getting used to crappy software, so they might not be surprised so much - and now I don't know if it is a good or bad news.

And sure I know that the modern 'nixes are fixed for this - but will all the systems which will be in service on that day be modern? And, more importantly, will the applications be updated for this? What's the value of a 64-bit time counter, if the application converts it into a signed 32-bit?

My point is slightly different, though. This is once again a famous example of crappy programmer's thinking. The shortcuts which we programmers do like to take. The pressure of "things can be done working in 1 day, why spending extra 5 days for making it bulletproof".

Well, it's weekend anyway, isn't it... :-)

JW


PS. Joe, my father manufactured several different products in the last 15 years he has his own business, but now, he (his employee) is sewing... ehm... underpants... :-)