Hacker News Reader

Comment by NikkiA | Hacker News Reader

6 digit years seem like a solution to a problem that will never be for the sake of looking forward thinking. There's absolutely no way current technology or social norms will last 8000 years.

View on HN

Yeah, but don't forget we use computers for things that don't involve today only.

Let's say you run a climate calculation for long enough for example. Now you could just say "fuck dates" if they error — but wouldn't it be kinda nice if they didn't?

If you want to project something beyond 9999-12-31, you can do that with custom timestamps, e.g. by specifying the year as a float. There's no need to convert that back to a timezone or sync it to UTC. The maximum difference of 1 day would be 0.00003% in 10K and decreasing, a level of accuracy no simulation can ever achieve.

Would you really need a full date and time stamp for doing that? If you're talking about projections 10,000 years in the future do you care if it's mid-afternoon on a tuesday?

It just makes great easter eggs.

"Hey, this guy had set his reminder to today. I wonder what he'd think if he knew we actually saw this 56,000 years later"

"Well, let's regenerate him and ask"

You can also use 5, or 7 digit years, or whatever the two parties can agree to before communication starts.

Part 2 of the standard gives examples of years with 10 digits.

6 digit years are a solution for a problem we have today. Think geologist simulating continental drift.

6 digit years aren't really _enough_ for that, tho; you're going to be looking at well over 100k years in most cases.

(I can definitely see cases where 6 digit years would be useful, tho)

Reply on news.ycombinator.com