> Note that the OpenVMS time display and manipulation routines allow for only 4 digits in the 'YEAR' field. We expect this to be corrected in a future release of OpenVMS sometime prior to 31-DEC-9999.
While they do have some active projects, like the 10K-Year Clock, it's mostly just a series of monthly-ish talks with a variety of ecologists, engineers, etc., themed around long-term thinking. It's all quite grounded.
it is pretty much cult-like, and the rabbit hole gets deeper the longer you look. highly recommend going to the related bar when/if you're in SF ever: https://theinterval.org/
Some might say the best way to preserve knowledge and culture for the long term would be to establish a successful cult. It seemed to work for the Egyptians and the Christians (experiment still ongoing). Just look at how we count years in the Western world.
My guess is because it seems to share people and terminology with some 60s and 70s counterculture and new-agey stuff.
I say that with total respect for said movements, but I see why cynical analysis would say cult-like. And there are a few people and groups in similar categories that are cult-like.
I read it as more of a PBS tote-bag / Patreon kind of thing. MLMs tend to promise fantastic riches, which LongNow certainly does not. (Not a member, but I've been following their talks for roughly 15 years now.)
They have just enough time to schedule meetings to discuss it, code it, review it, then deploy it. Except they'll discover on 15-DEC-9998 that there's a critical bug, and it'll be known as Y10K. All of the original developers will have been long gone, so there's no way of getting them involved. An entire new development environment will be created so they can run a client based language as a server to handle the new date format
Don't forget the 64-bit unix time epoch, that'll be a big one. 584 billion years from now, 2^64 seconds will roll over and just cause all sorts of mess.
Which means that 64-bit time_t is a somewhat-ultimate "lossless and future-proof archival format": you could use it to successfully round-trip-encode a reference to any historical or future second between the creation of our universe, and the last time life will be sustainable for carbon-based lifeforms in our universe.
Mind you, presumably, some civilizations in our universe will go on beyond that. Referencing this† chart, silicon-based life (e.g. maybe even us humans, if uploaded into computers) might be able to survive about half-way into "the Degenerate Era", before the material substrate we run on just radiates away through proton decay. Representing the last second life could exist would require a continuum of 4e32 seconds, or 108 bits.
And also keep in mind, computers (including any silicon-based life-forms) think much faster, and so are in want of much more fine-grained measurements than our unit of "seconds" to place events on their own timelines. We already use nanosecond timestamps for our computer-system-internal log events (which would add a 30-bit fractional part to our 108-bit timestamp, to create a combined 138-bit fixed-point representation.)
But let's be "safe." We don't know how fine-grained we'll need to get! We might want to record data about the first events of our universe, for example. I would think that a "really ultimate", future-proof, archival format for time record-keeping, would probably be denominated in Planck time units.
Interestingly, counting in Planck time units instead of nanoseconds, "only" requires 144 bits for the fractional part!
108 + 144 = 252. Or, rounding up, 256. We could represent any instant of time we could ever care about, in fixed-point encoding††, in a 256-bit word.
†† (It wouldn't really be fixed-point encoding. It'd just be an integer number of Planck time units since the Big Bang. But you can right-shift by 144 to get something close to a count of seconds, which is useful.)
Finally: you still want more? Any time between the creation and heat death of the universe, even though nothing will be around to experience it? Well, that'd be 333 bits worth of seconds. 333 + 144 = 477, which would fit a 512-bit machine word.
One could assume, then, that any Matrix-like universe simulation substrate will have a 512-bit Time Stamp Counter register.
...presuming, that is, that it was built to simulate a universe like our own, and not one that has finer-grained Planck units or which goes on for longer. :)
----
Does all of this have any use? Maybe! If we could figure out precisely what time it is in some objective sense—how many Planck time units have passed here in our reference frame since the creation of the universe—that might be a useful "handshake format" to describe "when" something happened to aliens, when building things like Voyager's Golden Record. (After all, we can teach them about our calendar, but how do we describe+transmit when our calendar began in a way they can measure against in their reference frame, in order to know how to do the math to translate dates back and forth?)
After all, we can teach them about our calendar, but how do we describe+transmit when our calendar began in a way they can measure against in their reference frame, in order to know how to do the math to translate dates back and forth?
Could this be done by sharing the distance between objects in space? No clue how accurately we could measure it, but if they could say hey year 0 was when earth was X LY from andromeda, and they can calculate how far they are now, and the rate of expansion, they might be able to calculate when year 0 was?
> One could assume, then, that any Matrix-like universe simulation substrate will have a 512-bit Time Stamp Counter register.
Maybe, though you can bet they don't have to deal with leap years and seconds. Let alone have to adjust the clocks as an earthquake slightly changed the speed of the rotation ever so slightly.
Makes you wonder what they would use as a reference point of time and what units of measure. Pulsars are great, but on the scale of the universe, maybe not that great.
Slow the definitional ticks down to be independent of (slowly but increasingly divergent from) Planck time and we can morph it to create as much time as we "need".
That's not actually true. Assuming there's a distinguished "initial" event [the big bang], and your universe does not contain any closed timelike curves [citation needed], you can define a Longest Proper Time from the initial event to any other event as the (obviously) longest proper time that any object or test particle could experience along a (timelike) path from the initial event to the later event.
Simultaneous events have equal LPTs and are always spacelike-separated (if there was a timelike path from A to B then B.lpt (>)= A.lpt + len(path) > A.lpt, so B.lpt!=A.lpt and they aren't actually simultaneous).
> An entire new development environment will be created so they can run a client based language as a server to handle the new date format
Is this in reference to something? Did some Y2K patches partially consist of putting Node.js-based gateways in front of mainframes to translate dates back and forth?
By 15-DEC-9998 Earth Corp will long be run by extra-terrestial hedge funds which forced theirs slav^employee on our planet to use their homeplanet's 2710.7-days-per-year timesystem...
> Note that the OpenVMS time display and manipulation routines allow for only 4 digits in the 'YEAR' field. We expect this to be corrected in a future release of OpenVMS sometime prior to 31-DEC-9999.