[conspire] UTC for the win! (was: Re: Ti-i-i-ime, is on my side, yes it is)

[Ross Bernheim]

> On Mar 11, 2020, at 8:13 PM, Michael Paoli <Michael.Paoli at cal.berkeley.edu> wrote:
> 
>> From: "Nick Moffitt" <nick at zork.net>
>> Subject: Re: [conspire] Ti-i-i-ime, is on my side, yes it is
>> Date: Mon, 9 Mar 2020 07:47:27 +0000
> 
>> If I were in charge, I'd just go full China and say we all use UTC and get used to what time things start/stop in our area.  No more arguing about what time it is
> 
> Here here!
> 
> Yep, were I the planet's time czar,
> we'd go full UTC planet-wide.
> https://en.wikipedia.org/wiki/Coordinated_Universal_Time
> 
> Other factors/benefits to keep in mind:
> Works really well across geography - including for portable compute
> systems that move.  Even works at the North and South Poles, and in space (
> well, for us Earth/human oriented folks/stuff).
> Ever deal accross timezones, or with logs of security related stuff?  If
> you ever deal with an external agency over such, many won't even look at
> one's logs unless they're UTC (most notably so they may be able to correlate
> events).
> 
> And if many/most folks really still want to do something
> Daylight/Summer Time-ish ... could just have, twice yearly,
> National "shift your schedule day" ... but dag nab it, do not
> screw with the clocks!  Sheesh!
> 
> Now, in the land/realm of POSIX, Unix, Linux, BSD, ...
> and leap seconds and such ... some minor bits to be worked out and/or
> squabbled over ... but if you start with UTC planet-wide, that gets
> even all that "exact", and within ~0.5s within 24 hrs. of a leap
> second event.
> 
> Now, there are arguments pro & con about how to deal with the leap
> seconds on Linux, etc. - and there are some different approaches ...
> but it would be good to much more standardize them ... so, say, we get
> it down to only and exactly at most 2 time disciplines ... and you're
> either on discipline one (exact time, insert leap second),
> or you go the smear route - keeps within 0.5s, but doesn't add the extra
> second as a distinct second ... but rather than umpteen or more
> different smears ... everyone uses the exact same one - planet-wide
> synced ... if they're doing the smear.  And, presuming it's not
> already in the NTP protocol, maybe they're ought be some flag that
> says, which discipline is being used, and flag to indicate currently
> active (smear in progress around leap second).
> 
> Personally I prefer using correct time, and adding the leap second, ...
> but unfortunately buggy software ... many prefer to go lower-risk
> by doing less accurate / less correct time, and fudging the leap
> second into a smear, so clocks are never off by more than 0.5s,
> and no minute ever has a 61st second.

I miss my atomic clock. Had to manually add or subtract leap seconds, but it wasn’t hard to do.

> And bloody heck, don't even think of changing to a leap minute!
> Egad, that's a crud idea.
> 
> 
> _______________________________________________
> conspire mailing list
> conspire at linuxmafia.com
> http://linuxmafia.com/mailman/listinfo/conspire