Hacker News Reader

Comment by egwor | Hacker News Reader

There's a significant cost to this work. Machines have to be upgraded: OS's changed, libraries updated. They did this in Russia and it caused us a headache due to python libs that needed updating and then OS changes. etc. etc.

View on HN

time zones change all the time. Most OS's automatically take these changes into account, I believe. ( see https://serverfault.com/questions/192858/updating-systems-ol... for a bit of info, but I think it is often included in OS updates rather than having to do it manually normally.)

https://en.wikipedia.org/wiki/Tz_database https://www.iana.org/time-zones

But of course there will be many legacy systems that assumes that all of Europe is on the same zone.

Taking the change into account for the current time is relatively easy, but date/time libraries in programming languages need to be updated so that you can write a calendar application with, say, a recurring event spanning across the change, or count days since a specific date in the past.

Even with DST, the actual date DST starts differs unpredictably from year to year. I don't think this introduces any new problems.

In the EU, the start date is predictable: the last Sunday in March. The end is equally predictable: the last Sunday in October.

Interesting.

I suppose that consistency could’ve given some EU only companies enough rope to hang themselves. Really, doing things the right way is easier than hand-writing some code that assumes DST starts/ends on a fixed date—at least in a language with a real date/time library.

At no point since the invention of timezones has Europe (the continent) been in a single timezone. Even if you only look at the EU, only time all of the EU was in the same timezone, was back in 1957 when "Belgium, France, Italy, Luxembourg, the Netherlands, and West Germany signed the Treaty of Rome, which created the European Economic Community (EEC)", and it wasn't even called EU. I would wonder which library makes that assumption.

Bad libraries that are written by people who think they don't need the tz.db to keep track of time, obviously.

Every year there are tens of changes to various time zones/calculations/all that fun stuff that get compiled into the tz updates that your OS vendor will ship and install.

This is nothing new.

Indeed time changes are a complex software issue, which is why getting rid of them and making it simpler is the long-term solution.

Because humans need to adapt to computers and not the other way around? And, sure, someone needs to program the libraries etc. that deal with timezones. But that's sort of why we program computers--to do what humans want done.

So first you complain that this will be too hard for computers to deal with because you'll need to update things. Then after it's explained that it's actually trivial for computers, you complain that it's too hard for people to manage and we shouldn't just do what's easy for computers. So which is it?

My intent was that computers (and programmers) need to adapt to how humans want to use systems. Not the other way around. I don't care if it's a PITA for programmers to deal with or not.

If they choose an existing timezone, not really, unless things aren't on NTP.

Just don't swing the tz over like normal.

Time Zones are already a huge PITA no matter how you slice it.

This video sums it up fairly nicely: https://www.youtube.com/watch?v=-5wpm-gesOY

That sound you can faintly hear is the screaming of a million developers who maintain all the datetime libraries in the world.

Reply on news.ycombinator.com