Hacker News Reader

Comment by netcraft | Hacker News Reader

some info to save someone else from having to look it up

  1 400 000 000 was 5/13/2014, 11:53:20 AM
  1 600 000 000 will be 9/13/2020, 7:26:40 AM
  1 000 000 000 was 9/8/2001, 8:46:40 PM
  2 000 000 000 will be 5/17/2033, 10:33:20 PM
  2^32-1 (4 294 967 295) will be 2/7/2106, 12:28:15 AM

  100 million seconds is a little over 3 years
  1 billion seconds is almost 32 years.

View on HN

You have skipped over one interesting date between 2e9 and 2^32-1 - 2038-01-19T03:14:07:

https://en.wikipedia.org/wiki/Year_2038_problem

Although perhaps you have assumed that everyone already knows it!

I've been programming for a billion seconds since the timestamp was in 500 M. Good ol days with Turbo Pascal and Clipper Summer 85, now it's all about Swift and Kotlin.

Time flies.

I remember 1e+9 because it shifted a log file over by 1 character. And less than a week later I was out of a job (travel industry).

Are these events related to each other?

Was that a coincidence? Please don't leave us hanging.

1 billion Unix time was September 8, 2001.

Sep 9th actually. 0146 GMT.

Don't you see? Whether one event led to another is not important. We don't need to know for sure. It's like the end of Inception.

Log file break/fix did not result in termination. But I probably wouldn't have remembered the former if it weren't for the latter.

I've conjectured before that 9-11 happenening so close to the epoch flip was deliberate, as a hint that middle-east radicals weren't really disorganized, saber-wielding incompetents as Hollywood portrayed them.

What do you think are other interesting upcoming timestamps? - http://timestamp.online/countdown/1515151515 - http://timestamp.online/countdown/1579119751

http://timestamp.online/countdown/1610612736

Obviously. Programmers celebrating 1500000000 just make me sad.

As I mentioned in the Reddit thread, everyone is looking at decimal but overlooking the (arguably more interesting) base-16 dates.

Jan 13 19:13:30 2018 UTC will be 0x5A5A5A5A

0x66666666 will be Jun 10 2:35:18 2024 UTC

We just missed a good one:

  Sep 22 2016 16:00:00
  in base10 1474560000
  in base16 0x57E40000

See the answers to the earlier time that you asked this question.

* https://news.ycombinator.com/item?id=14759814

And my favorite, π seconds is a nanocentury.

How so?

I had assumed there would be an error in there about forgetting that years don't have an integral number of days, but even at exactly 36500 days per century, a nanocentury is still over 3.15 seconds.

nano century = 100 * 365.25 * 24 * 60 * 60 * 1e-9 = 3.15576

(nano century - π)/π = 0.45%

So a nano century is equal to π seconds within half a percent which is more than good enough for casual uses as an approximation.

It's also equal to 3 seconds within an error of 5.2%, which is also more than good enough for casual uses.

But when you're having fun with pi, you generally measure similarity in terms of being accurate to some number of decimal places, and this statistic is accurate to an unimpressive one place (3.1).

Well one second is a microfortnight

Also 2038-01-19 https://en.wikipedia.org/wiki/Year_2038_problem

timestamp is stored as a signed value since it can also go backwards from the epoch, so you need 2^31-1

Reply on news.ycombinator.com