Was the Y2K scare a huge scam by computer companies?

Bug expert wrote:Y2k was largely, though not exclusively, a Cobol thing.

My dad was a cobol programmer in the 60s and 70s. The Y2K hype drove him up the wall and he was writing letters to all the scaremongers, which were doomsday prophets, not computer companies. He kept saying that they knew about the issue back in the 60s but figured it'd be fixed by 2000, that it wasn't that hard to fix, and that there'd be problems but most of them minor if it wasn't fixed.

a day late... wrote:

Bug expert wrote:Y2k was largely, though not exclusively, a Cobol thing.

My dad was a cobol programmer in the 60s and 70s. The Y2K hype drove him up the wall and he was writing letters to all the scaremongers, which were doomsday prophets, not computer companies. He kept saying that they knew about the issue back in the 60s but figured it'd be fixed by 2000, that it wasn't that hard to fix, and that there'd be problems but most of them minor if it wasn't fixed.

The old Cobol programmer I knew was very happy making $200 / hour, 12 hours a day, 6 days a week fixing the Y2K problem. Good times.

DiscoGary wrote:

Bug expert wrote:

2038 bug is going to be way bigger.

Y2k was largely, though not exclusively, a Cobol thing.

Unix time will overflow sign 32 bit integers in 2038, and virtually every computer program in the world relies on Unix time. Virtually all of them will be running on 64 bit hardware by then, but there are millions of databases that will have to have their schemas migrated.

He's referring to the 1970 based "epoch" where some computers calculate time based on the number of seconds since Jan 1, 1970. In the year 2038 that number will exceed 2^31 which means any computer still running in 32 bit mode will think time has been reset to 1970.

We don't have to worry about that because Google robots will have eradicated the human race by then.

I used to think Google was Skynet until they changed their name to Umbrella corporation. So now I'd have to say that they will eradicate the human race, but with zombies, not robots.

DiscoGary wrote:

Bug expert wrote:

2038 bug is going to be way bigger.

Y2k was largely, though not exclusively, a Cobol thing.

Unix time will overflow sign 32 bit integers in 2038, and virtually every computer program in the world relies on Unix time. Virtually all of them will be running on 64 bit hardware by then, but there are millions of databases that will have to have their schemas migrated.

He's referring to the 1970 based "epoch" where some computers calculate time based on the number of seconds since Jan 1, 1970. In the year 2038 that number will exceed 2^31 which means any computer still running in 32 bit mode will think time has been reset to 1970.

We don't have to worry about that because Google robots will have eradicated the human race by then.

You're understanding of the 2038 bug is incorrect. The computers will not think that it is 1970, they will think it is 1901. That is because they're using signed 32 bit, which means the systems will read a negative number when it all "resets".

I have a grand plan wrote:

DiscoGary wrote:

He's referring to the 1970 based "epoch" where some computers calculate time based on the number of seconds since Jan 1, 1970. In the year 2038 that number will exceed 2^31 which means any computer still running in 32 bit mode will think time has been reset to 1970.

We don't have to worry about that because Google robots will have eradicated the human race by then.

You're understanding of the 2038 bug is incorrect. The computers will not think that it is 1970, they will think it is 1901. That is because they're using signed 32 bit, which means the systems will read a negative number when it all "resets".

Yup. That is correct.

But I notice you didn't dispute my prediction that Google robots will have wiped us all out by then, so I got at least half of it right.

DiscoGary wrote:

I have a grand plan wrote:

You're understanding of the 2038 bug is incorrect. The computers will not think that it is 1970, they will think it is 1901. That is because they're using signed 32 bit, which means the systems will read a negative number when it all "resets".

Yup. That is correct.

But I notice you didn't dispute my prediction that Google robots will have wiped us all out by then, so I got at least half of it right.

Definitely not disputing that. We're f*cked.

Lenny Leonard wrote:

fgfg wrote:

back when "loggin on to the internet" was shitting your swim trunks

Simpsons joke between my handle and Carl Carlson.

Golf Clap for you.

https://www.bloomberg.com/news/articles/2017-06-15/trump-orders-government-to-stop-work-on-y2k-bug-17-years-later

What seems odd to me is that we spent billions to avert disaster, but other countries spent zero and had no problems. I've never heard of any incident anywhere. Not so much as someone's Amiga computer needing to be restarted. I'm curious to hear from the Y2K experts why there were no problems in places that did nothing.

My dad was one of those programmers in the 60s with the white shirts and black glasses. The Y2K panic drove him up the wall. He was writing letters to media outlets telling them we've known about it for years, we'll fix it, and even if we don't nothing catastrophic will happen. He was demanding they give public apologies for scaring people. Needless to say, his letters were ignored.

I was a COBOL programmer for about 6 months in 1984-5, and we were using four-digit year fields then. I remember being told that at one point (1960s?), memory (or something) was so limited/expensive that programmers didn't use four-digit year fields if they thought their code wouldn't be in use by 2000 or ever need to work with post-1999 years, but that it had become so much cheaper that everyone was using four-digit fields.

Bug expert wrote:

2038 bug is going to be way bigger.

Y2k was largely, though not exclusively, a Cobol thing.

Unix time will overflow sign 32 bit integers in 2038, and virtually every computer program in the world relies on Unix time. Virtually all of them will be running on 64 bit hardware by then, but there are millions of databases that will have to have their schemas migrated.

I'd be more worried about all those things connected to the internet of things.

JamesD wrote:

I was a COBOL programmer for about 6 months in 1984-5, and we were using four-digit year fields then. I remember being told that at one point (1960s?), memory (or something) was so limited/expensive that programmers didn't use four-digit year fields if they thought their code wouldn't be in use by 2000 or ever need to work with post-1999 years, but that it had become so much cheaper that everyone was using four-digit fields.

This is exactly right. My dad was a COBOL programmer in the 60s (started with machine language) and he has stories of desperate appeals for the company to spend a fortune to get an extravagant increase of 1k of memory. Like you said, they were aware of the Y2K issue back then but figured their code would be out of use or updated sometime before 30 years.