@HDE226868 Then the obvious follow-up question is this: how can this be mitigated? Because — of course — the users of this time reference do not need to sit still on their arses — or whatever counts for them — and shrug their shoulder-analogues and go "Okay... guess that's that... back to flying clocks around". If there is an inherent flaw with using pulsars, then there should be ways around it, to minimise the accuracy error. Such as: use more pulsars for reference; update the catalogue frequently enough; use local "tickers" to detect glitches and to compensate for drift.

Also, I am not looking for ideal — or perfect — here... I am looking for good enough. Remember the old maxim: Perfection is the natural enemy of Good Enough.

Pulsar clocks have been suggested, true, but only as a replacement for the balance wheels in wristwatches. I have downvoted this answer because it attributes magical powers to pulsars. No, hearing a clock going tick-tock-tick-tock and another clock going tick...tock...tick...tock does not enable you to tell time neither here on Earth, nor anywhere else. Those "precise clocks" mentioned in passing in the Weaky Paedia article are supposed to work by first setting them to a precise time and then counting the pulses. See the problem?

No @AlexP, I really do not see the problem. Because if you hear tick-tick-tick from enough sources, and you know their respective period, and you know what their relative phases to each other were at a given time... then you do know what time it is. It is a trivial problem to calculate it. Also... I am kind of curious what you believe the use of a derogatory name for WP is meant to showcase... other than your personal bias and lack of arguments.

@HDE226868 That is why I set the tolerance really wide... a whole second, give or take, which in time-metrology terms is a huge amount. The catalogue and the time standard must have been disseminated at some point. And for as long as there is a need to have synchronised clocks, there must also be a way to disseminate updates to that time reference.

"You know what their relative phases to each other were at a given time" ... where? The phases of the tickings will be different in different frames of reference mving at speed chaotically, and they will vary in time. (Not to mention the rather difficult problem of defining phase between two or more signals with different periods of oscillation.)

No, @AlexP, it is not a difficult problem. And no, the point of pulsars is that they do not behave chaotically but very predictably. Just because you do not know the math for how to do this, or because you use ridiculous names like "Weaky Pedia", does not mean everyone else is as lost as you on the matter. If you have a better suggestion, then by all means present it as an answer. But this line of commenting of just trying to disparage another answer for no better reason than — in effect — "I don't get it", is abjectly pointless. I advise you to find something better to do with your time.

@AlexP No, not a difficult problem... for those that know the math, which you do not, apparently. Essentially it is a problem very close to that of calculating your time and position via GPS. Slightly more tricky since you get ticks here and not the time stamps, but, never the less doable.

@MichaelK: The key aspect of the math which I don't know is that you need to know the position of the signal sources with great precision. For a signal with a period of about 30 msec, an error of 10,000 km in the position of the signal source will make the phase of the signal completely indeterminate. How do you propose to measure distances, masses, and velocities with sufficient precision to make a sufficiently accurate map over many millennia?