Clock synchronization is a nightmare • Hacker News Board

Re: Clock synchronization is a nightmare

#161

Post by Dylan16807 » Tue, Dec 30, 2025, 1:28 AM UTC

kobieps wrote: post #156

Dylan16807 wrote: post #109

Earlier quoted context omitted.

I would not call "loses track of time if it's [partially] unplugged" a nightmare.

Haha fair, but in this case it was "loses the time if the power supply is interrupted"

I'm counting power loss as a type of being unplugged for that loose statement, even if technically they're different.

Re: Clock synchronization is a nightmare

#162

Post by hdjrudni » Tue, Dec 30, 2025, 11:46 PM UTC

varjag wrote: post #133

8n4vidtmkvmk wrote: post #118

Earlier quoted context omitted.

If it's for an event, can they not bring all the devices together in close proximity and sync them somehow? That at least removes network delays

You can't sync individual oscillators precisely for very long.

Not even if you test hundreds of pairs to find a match?

Re: Clock synchronization is a nightmare

#163

Post by baby_souffle » Wed, Dec 31, 2025, 9:44 PM UTC

hdjrudni wrote: post #162

varjag wrote: post #133

Earlier quoted context omitted.

You can't sync individual oscillators precisely for very long.

Not even if you test hundreds of pairs to find a match?

> Not even if you test hundreds of pairs to find a match?

Assuming you do find _a_ match, you still need everything else to be in sync across the different temperature(s) that each component will be operating at

Re: Clock synchronization is a nightmare

#164

Post by shomp » Fri, Jan 02, 2026, 4:52 AM UTC

Enginerrrd wrote: post #158

tobias2014 wrote: post #106

Earlier quoted context omitted.

For any space-like event you can find reference frames where things happen in different order. For the time-like situation you described the order indeed exists within the cone, which is to say that causality exists.

You can still order them with the spacetime interval compared to a reference event, even for space like separated events. It allows for differing elements of the set to share the same value but so does using time alone. It just also allows every observer to agree on the ordering. Bc Assigning a distance function to elements of a set is a common way to do that in fact. It doesn’t work with just a time coordinate or sp…

I think you meant compared to a reference observer? Events are not really independent of observers. Consider the case in baseball where a runner and the baseman tag the base at the "same" time from opposite sides of the base. Assume they move at equal speeds. If the umpire is closer to the baseman then the baseman has tagged it first, if he is closer to the runner, then the runner has tagged it first. The "event" of "touching the base" has two possible outcomes depending on where the observer stands, and there is no "view from nowhere" or observer-free view that we can reference.

Re: Clock synchronization is a nightmare

#165

Post by Enginerrrd » Wed, Jan 07, 2026, 3:00 PM UTC

shomp wrote: post #164

Enginerrrd wrote: post #158

Earlier quoted context omitted.

You can still order them with the spacetime interval compared to a reference event, even for space like separated events. It allows for differing elements of the set to share the same value but so does using time alone. It just also allows every observer to agree on the ordering. Bc Assigning a distance function to elements of a set is a common way to do that in fact. It doesn’t work with just a time coordinate or sp…

I think you meant compared to a reference observer? Events are not really independent of observers. Consider the case in baseball where a runner and the baseman tag the base at the "same" time from opposite sides of the base. Assume they move at equal speeds. If the umpire is closer to the baseman then the baseman has tagged it first, if he is closer to the runner, then the runner has tagged it first. The "event" of…

No, I mean a reference event, though you bring up an interesting subtlety. (Essentially I just mean an event that definitely happened [A particle decay, a supernova, an omnidirectional radio signal, etc] which will serve essentially as an origin point on the spacetime manifold). You are right though that technically, we need at least one observer to define the coordinates of that event initially. Once that's done however, ALL observers can order events according to the spacetime interval between any event they observe and the reference point (transformed into their coordinates) and they will ALL agree on that ordering. A "good" event here would be something that observers can compare. I think using pulsar pulses counted from some epoch is a perfectly good reference here, assuming we could communicate that omnidirectionally. The difference, as measured by the spacetime interval, between any event in any observers reference frame, and a reference event in their past lightcones is something that ALL observers that can communicate will always agree on. Observers may disagree about how many pulses have occurred since that epoch at a particular time in their coordinate time, but it doesn't matter. As long as they're comparing in spacetime intervals to a particular count on the pulsar, no disagreement will occur. i.e. the spacetime interval between the 3rd pulse and some event will always be the same since it's a lorentz invariant scalar quantity (i.e. a rank zero tensor).

Your baseball analogy has flaws: No properly defined "event" in spacetime will have dual-outcomes. The events in that case are that "a baseman tagged the base", and "a runner tagged the base". "x tagged the base first" is NOT an event, that's a comparison between events, and it's one that was done in a particular observers time coordinate, which is not the correct procedure here. No Lorentz invariant transformation between observers within the light cone will disagree that those events happened, though observers may disagree which happened first within their coordinate time.

(Note the issue of observers needing to be in the same light-cone is a superficial one. I haven't defined that precisely, but I don't need to: If observers can communicate at all, they will agree, upon communication, that an event is within their past light cone. In the context of server synchronization, this will always be true.)