Talk:Timeline of the far future/Archive 3

11th millennium and beyond
I've kept a stump of the material from that article that I haven't been able to cite and move here yet. If anyone can cite or refute the information on this list, that would be much appreciated.  Serendi pod ous  17:24, 6 August 2012 (UTC)
 * I'm not sure that any of them deserve to be included based on WP:WEIGHT. Shrug. Regards, RJH (talk) 19:38, 6 August 2012 (UTC)

Is this a valid external link?
People don't seem to get Poincare recurrence, so I thought adding this video would be useful.  Serendi pod ous  07:50, 7 September 2012 (UTC)

Earth life more durable?
A new calculation indicates extremophile microbes may last till 2.8 billion years in polar caves. — Preceding unsigned comment added by 174.100.29.249 (talk) 18:41, 31 October 2012 (UTC)
 * Excellent ref. Thanks for that.  Serendi pod ous  09:32, 1 November 2012 (UTC)

End of Current Interglacial
Now the article reads

"50,000 years from now. According to the work of Berger and Loutre,[7] at this time the current interglacial ends, sending the Earth back into an glacial period of the current ice age, assuming limited effects of anthropogenic global warming."

But here there is another rather different prediction:

"The glacial inception during Marine Isotope sub-Stage 19c, a close analogue for the present interglacial, occurred near the summer insolation minimum, suggesting that the interglacial was not prolonged by subdued radiative forcing.[7] Assuming that ice growth mainly responds to insolation and CO2 forcing, this analogy suggests that the end of the current interglacial would occur within the next 1500 years, if atmospheric CO2 concentrations did not exceed 240±5 ppmv."

— Tzedakis, P.C. et al. Determining the natural length of the current interglacial. Nature Geoscience (2012) vol. 5 (2) pp. 138-141 http://www.nature.com/ngeo/journal/vaop/ncurrent/full/ngeo1358.html

80.174.254.239 (talk) 18:54, 27 November 2012 (UTC)
 * Perhaps, but that estimate is too soon for this list; the entrance point is 10,000 AD +  Serendi pod ous  23:50, 27 November 2012 (UTC)


 * How about adding a note mentioning that a different model predicts a possible end of the interglacial for much earlier, although this result is only expected if CO2 atmospheric concentration remains below certain level?


 * I've had a look at the timeline of the near future and it extends only until the end of the 23rd century (until the decade of the 2290s), 1,500 years would be 5 times as much. On the other hand, the timeline of the near future has been divided into decades so adding an entry for an imprecise event in the net 1,500 years would break the current arrangement, whereas in the timeline of the far future this compartmentalization doesn't exist. 80.174.254.239 (talk) 04:46, 28 November 2012 (UTC)
 * Because this list goes beyond numbers we even have names for, such compartmentalisation would be impossible. I am sure that there are many, many models predicting the return of the next glacial maximum. Why cite just one alternative? Of course such speculations are tentative at best; that's why I cited the names of the authors in the list itself. As far as a place to put your information goes, right now, that would be 4th millennium.  Serendi pod ous  09:18, 28 November 2012 (UTC)

Minor self-contradiction
In the "Astronomical events" section, in this row: "The Julian calendar (365.25 days) and Gregorian calendar (365.2425 days) will be one year apart." The date given in text (1 March, 48,901 AD) is different than the one the time delta is calculated from using age in years and days (1 April). Somebody should check the math there. (I noticed this when translating the article.) Matma Rex pl.wiki talk 11:20, 30 November 2012 (UTC)
 * Well spotted! :-)  Serendi pod ous  12:31, 30 November 2012 (UTC)

Voyager 2/Ross 248
This is a decent contribution; however it seems to grate against the other data; why would it take Voyager 2 ten times faster to pass Ross 248 as Voyager 1, particularly when it is travelling more slowly? I cannot read the specific page of the cited book online so I don't know if the fault is with the source, the editor or with the other cited NASA pages. But I would like a direct quote, preferably by someone other than the original poster, before I'm satisfied. Thanks.  Serendi pod ous  02:53, 4 February 2013 (UTC)

"Barring an unlikely collision, the craft should persist indefinitely"
I remember reading there's friction in space, from tiny particles scattered the universe, that will eventually, over a very long distance, slow down to a halt anything in space. I guess indefinite can fit, but isn't there a better way to phrase it? -- occono (talk) 15:15, 9 March 2013 (UTC)
 * The interstellar medium? I don't have any calculations but I think it would take many, many, many, orders of magnitude more time then the 8 million years listed in that section for it to have any sort of effect on velocity. Anyway, the page said the craft would persist, i.e. to continue to exist, not remain in motion, indefinitely.  Reatlas  (talk) 15:52, 9 March 2013 (UTC)
 * That would need a source. The cited comment is sourced.  Serendi pod ous  19:21, 9 March 2013 (UTC)
 * I misread it, nevermind :P -- occono (talk) 20:59, 9 March 2013 (UTC)

Predictions
I think the reason why someone added a fact tag to While predictions of the future can never be absolutely certain might be that they mistook certain for accurate. The phrase doesn't say the predictions are all wrong, it says that we can't know if they'll turn out right or wrong. &mdash;&thinsp; H HHIPPO  19:28, 6 May 2013 (UTC)

Fact
Ok. So this is the first sentence: "While predictions of the future can never be absolutely certain.."

I believe this sentence need citation. How can you be so sure that predictions of the future can never be absolutely certain? How do you know? If it is a fact then give some sources. I am going to revert it back. Please discuss it here before removing it without any reason. — Preceding unsigned comment added by 95.6.27.202 (talk) 17:48, 10 May 2013 (UTC)
 * And... you're wrong. Nothing in life is absolutely certain unless it already happened. In the next ten minutes, the world could end. Or, we could turn into pizzas. Or, nothing happens. hence, here is my reason. That sentence is factually and logically correct.  HHHIPPO's explanation at least sounded plausible. Your explanation is exactly what I thought it would be, and that's why I removed it. And I gave you a reason. I have to assume that your edits are made in good faith. But, they sure don't seem like it. — Preceding unsigned comment added by Kude90 (talk • contribs) 19:37, 10 May 2013 (UTC)
 * Being a featured list does not excuse this article from needing sources. If it's factually and logically correct then you should have no problem finding a reliable source for this statement anyway. Simply giving your reason and insisting it's right without a independent RS is original research and it isn't sufficient.  Reatlas  (talk)  01:55, 11 May 2013 (UTC)
 * That is exactly the my point. You are talking very sure about it. And my question is still same: How can you be so sure that predictions of the far future can never be absolutely certain? Please provide at least one reliable source which point out scientifically that predictions of future is impossible. Who knows, people might develop a new technology in the future which helps to predict the future. And one last thing; i am admirer of this well written article and i am doing my edits in a good faith. Just because i have no account here doesnt mean i have no good faith in my edits. — Preceding unsigned comment added by 95.6.27.202 (talk) 06:15, 11 May 2013 (UTC)
 * Well, how could you be absolutely certain that that technology is working as you think? I think this is a rather philosophical discussion, and we could spend ages on it only to find out that we disagree on the definition of 'certain'. Alternatively, we could discuss if there are other reasons for removing or modifying the statement, for example:
 * The whole article is not about philosophy, so the first sentence in the lede shouldn't be either.
 * The statement "predictions of the future can never be certain" is actually not needed for the storyline of the article, all wee need is that there are no certain predictions now.
 * Maybe changing "can never be" to "aren't" would be an option, I'm not sure that's the best solution though. &mdash;&thinsp; H HHIPPO  10:24, 11 May 2013 (UTC)
 * I don't have time to explain to the OP the philosophy of inductive reasoning or the discoveries of Werner Heisenberg or Kurt Godel. Suffice to say, no, we can never be absolutely certain about the future.  Serendi pod ous  15:32, 11 May 2013 (UTC)
 * Good points indeed. Still it would be better with a reference, so here's one:
 * Note that an IP has today removed the statement in question from the article. To do this in the middle of an ongoing discussion is rather disruptive, but I'll refrain from reverting for now. But I do suggest we return the statement (with citation) soon. &mdash;&thinsp; H HHIPPO  10:12, 12 May 2013 (UTC)
 * Thanks. Ref added.  Serendi pod ous  16:37, 15 May 2013 (UTC)
 * Thanks. Ref added.  Serendi pod ous  16:37, 15 May 2013 (UTC)

Statistical probability events
Just a minor suggestion: Some of the listed events likely to have occured within an interval of time. Wouldn't it be a good idea to more clearly discern between approximations of an event, say Ross 248 that is expected to occur around the mentioned date, and an event that might happen at any time although having a high probability of having happened between now and the stated future date? It is indeed mentioned in varying ways in the describing text, but it seems somewhat ambigous and some might interpret that as an event probable to happen around the mentioned future date. I am thinking something along the line of; having a standardised syntax at the start of the description for these events or; giving a time interval instead of a date, i.e. 1 - 1 million or; having a symbol like an asterisk or a custom one to denote the distinction. Thoughts? — Preceding unsigned comment added by Ozmonatov (talk • contribs) 22:21, 12 August 2013 (UTC)
 * Makes sense to clarify, but instead of an asterix/symbol, I suggest using a footnote. E.g.
 * —  Reatlas  (talk)  08:07, 13 August 2013 (UTC)
 * Note added; I'm a bit unsure about 3.3 billion years; obviously that's a statistical guess, but I'm assuming that, even if it is, it still falls outside the 230-million-year planetary Lyapunov time.  Serendi pod ous  09:02, 13 August 2013 (UTC)
 * Note added; I'm a bit unsure about 3.3 billion years; obviously that's a statistical guess, but I'm assuming that, even if it is, it still falls outside the 230-million-year planetary Lyapunov time.  Serendi pod ous  09:02, 13 August 2013 (UTC)

While there's discussion going on, I may as well bring this up again: Should I list this as an external link? A lot of people aren't getting the Pointcare recurrence thing, including me, and this explains it better than the article does.  Serendi pod ous  12:42, 13 August 2013 (UTC)


 * Can't hurt. The current Poincaré recurrence time listings are at the best rather complex and abstract, and at worst verging on incomprehensible technobabble. I confess, video or not, I still don't get it. —  Reatlas  (talk)  13:28, 13 August 2013 (UTC)

Question for admins
I've been thinking of adding this edit notice:

'Jokes about the release of Half-Life 3'' have been added to this page at least five times already. Doing so again will not make you look smart, only unimaginative.'''

I think, phrased like that, it may deter future additions. Phrased any other way, it may actually encourage them.  Serendi pod ous  08:59, 20 August 2013 (UTC)


 * Not an admin, but there are two ways I can see in which this might be counterproductive. 1) WP:DENY recognition. Adding a message is an admission that they are successfully disrupting Wikipedia, and insulting them shows people are getting angry/frustrated. 2) WP:BEANS. Reminding people not to add such a joke can tempt them to do it. —  Reatlas  (talk)  09:53, 20 August 2013 (UTC)

Source 13
The statement is made that "Earth will likely have undergone a supervolcanic eruption large enough to erupt 3,200 km3 of magma; an event comparable to the Toba supereruption 75,000 years ago." is unsubstantiated as its own source is nonexistent. Perhaps it may or may not have existed in the past. Since it is not uncommon for an amateur or student to incorrectly interpret an article and write the wrong information about a topic I would at most list this statement as circumspect. — Preceding unsigned comment added by 2601:B:A3C0:7:A132:A846:53EF:24D0 (talk) 22:34, 27 August 2013 (UTC)
 * Woah. Ease up dude. Web pages go dark all the time. Thanks for the headsup. Anyway, found the original cited pdf for the information, so everything's cool.  Serendi pod ous  08:28, 28 August 2013 (UTC)

Decay
The article claims that years is the "Estimated time until the Earth's orbit around the Sun decays via emission of gravitational radiation", referencing it with http://rmp.aps.org/abstract/RMP/v51/i3/p447_1, to which I have no access. What exactly does "decays" mean here? Orbital decay is a long-running process. Does this imply that this is the time when Earth and Sun will collide? Or will the effect merely become noticeable?

(There is another similar statement at 50 billion years, "Thereafter, the tidal action of the Sun will extract angular momentum from the system, causing the lunar orbit to decay and the Earth's spin to accelerate.", with similarly ambiguous meaning; the source there is available, but I couldn't find where it mentioned anything about 50 bln years in particular, so I assume that the "thereafter" part makes this valid.)

Matma Rex talk 17:21, 8 September 2013 (UTC)
 * Clarified.  Serendi <sup style="color:#bb0000;">pod ous  11:13, 21 September 2013 (UTC)
 * Thank you! You might want to know that with this I have essentially finished my translation of this article to Polish (draft: pl:Wikipedysta:Matma Rex/Timeline of the far future) and I am intending to run it through the featured article process there (I will relay any comments relevant to the original version here) :). Matma Rex talk 00:09, 22 September 2013 (UTC)

Black Dwarf
According to the timeline, the Sun becomes a white dwarf at 8 billion years. At 14.4 billion years, it becomes a black dwarf. That is an elapsed time of only 6.4 billion years, while the black dwarf page the item links to says "Because the time required for a white dwarf to reach this state is calculated to be longer than the current age of the universe (13.8 billion years), no black dwarfs are expected to exist in the universe yet..." The source cited for the item does propose that it's possible for black dwarfs to evolve in only 7 billion years, but it is referring to stars of only 0.6 solar masses. Furthermore, the 14.4 billion years item claims: "[the sun's] temperature falls to 2239 K, making it invisible to human eyes." Objects such as brown dwarfs have temperatures cooler than 2239 K, and that does not mean they are invisible. Possibly whoever wrote the item meant the sun would no longer radiate visible light and would therefore no longer be visible from Earth to the human eye, but there's nothing to back that up in the cited reference EricJamesStone (talk) 19:07, 26 October 2013 (UTC)

Pioneer 10 passes within 3 light years of Ross 248
"Pioneer 10 passes within 3 light years of Ross 248", but in 300,000 years or just 30,000? The source is dead, but googling points to 30k; it's also more consistent with "The small red dwarf star Ross 248 passes within 3.024 light years of Earth, becoming the closest star to the Sun." in 36,000 years in the first section. Do we have a better source for this than, say, APOD? (This was spotted by Happa during the pre-DYK review of a translation of this article to Polish at .) Matma Rex talk 19:52, 22 December 2013 (UTC)
 * NASA science news also says 300,000. There are a couple other sources that say 32,600, but these aren't as reliable. Will need some form of direct confirmation.

EDIT: I've reinstated the number with another citation. I am fully aware that the citation may be incorrect; NASA have committed some howlers in the past on their site, though usually this was astronomical information, which can be corrected with better sources. In this case, these are probes that NASA itself launched. Any other source is likely to get their information from them.

EDIT: OK, I've scanned a number of nasa.gov sites and the majority do appear to give a 30-32,000 year number. While I have no evidence to make the call I can go with the majority. This looks like another NASA howler. Sheesh. The number of times I've come a cropper referencing that fricking site...

 Serendi <sup style="color:#bb0000;">pod ous  20:11, 22 December 2013 (UTC)

Thank you Very Much BBC Worldwide /sarcasm
Can someone who lives outside the UK please see if their recently posted infographic http://www.bbc.com/future/story/20140105-timeline-of-the-far-future‎ properly credits Wikipedia? Would be nice if we could get citations for some of their additions too.  Serendi <sup style="color:#bb0000;">pod ous  12:22, 9 January 2014 (UTC) Serendi <sup style="color:#bb0000;">pod  ous  08:30, 10 January 2014 (UTC)
 * Not sure why it would need to credit Wikipedia, but to answer your question, no, the infographic doesn't appear to credit or mention Wikipedia. Rolf H Nelson (talk) 04:24, 10 January 2014 (UTC)
 * Plagiarism is plagiarism, wherever you get it from.  Serendi <sup style="color:#bb0000;">pod ous  08:30, 10 January 2014 (UTC)
 * What part is plagiarized? For example, are there parts where the exact wording was the same? If there's no clear proof, Wikipedia is probably likely to let it slide (glass houses and all that.) Rolf H Nelson (talk) 22:25, 11 January 2014 (UTC)
 * I'm aware that the case is fairly minor, but yes, there are several snatches of text from this article. I'm really just pissed off because I wrote most of them.  Serendi <sup style="color:#bb0000;">pod ous  22:28, 11 January 2014 (UTC)

Boltzmann brains
I also think we should delete the Boltzmann brain item cited to Andrei Linde. The discussion that Boltzmann brains outnumber normal brains in many eternal-inflation models is interesting and probably even deserves its own page, but Linde appears to claim that axiomatically Boltzmann brains can *not* in fact outnumber normal brains, and therefore does not seem to actually be claiming Boltzmann brains will appear at the given timeframe. In other words, Linde and others appear to be using the BB timeline as a reductio ad absurdum to rule out certain theories, rather than seriously proposing that the BB timeline is actually correct. Rolf H Nelson (talk) 04:20, 10 January 2014 (UTC)
 * The paper says "the typical estimate of the rate of the BB production of 1B ∼ 10^10^50"; the word "typical" suggests that number has been arrived at by others. As such, the paper's opinion of its feasibility is irrelevant, particularly given that most the ideas we have about the far future are speculative anyway, and alternate futures are already mentioned on this list- see the fact that a new Big Bang occurs before the Boltzmann Brain.  Serendi <sup style="color:#bb0000;">pod ous  08:38, 10 January 2014 (UTC)
 * Can you clarify? According to the current timeline, a Boltzmann brain occurs long before a Big Bang. I can find sources that seriously propose that new big bangs may actually happen, but I haven't found a source that says a Boltzmann Brain will actually happen, or more importantly that such occurence has significance beyond being a Reductio ad Absurdum. Rolf H Nelson (talk) 22:16, 11 January 2014 (UTC)
 * The paper cites the number as one that is typically cited by other writers. The only thing that this timeline notes is the number and the formation of a Boltzmann brain. There is no guarantee that any of the events in this timeline will actually happen, so I don't really see the issue with such a hypothetical. The number has nothing to do with the paradox itself, which states that such entities are far more likely than evolved brains (and I think is nonsense); it's simply how long it would take for one to form, probabilistically speaking.  Serendi <sup style="color:#bb0000;">pod ous  22:40, 11 January 2014 (UTC)

Poincare recurrence
I think we should delete the three "estimated Poincaré recurrence time" items cited to Don Page. Don Page's calculations don't seem to yet be considered significant by any other source to my knowledge (though I could be wrong), and Page himself introduces the calculation with "One can also amusingly calculate that..." Given that most of the other items are much better sourced, I don't think this belongs on this particular list. Rolf H Nelson (talk) 04:14, 10 January 2014 (UTC)
 * The paper itself is weird, but the calculation is not all that controversial; it's pretty simple. He's just the only one who's bothered to do it.  Serendi <sup style="color:#bb0000;">pod ous  08:27, 10 January 2014 (UTC)
 * I'll accept that the calculation is simple. However, my concern is that I don't see why the calculation is significant, as it has much less WP:WEIGHT than the other items on the list. In the absence of sources, I don't see how the "Poincaré recurrence time for the quantum state of a hypothetical box containing an isolated black hole of stellar mass" has any direct physical significance, as no such system can exist long enough to recur without being dispersed by our Universe's expansion. It could have significance if you say that you're talking about a vaccuum fluctuation producing such a system, but now we're getting to where the reader would have to read the original paper or we'd have to add inline several paragraphs of original text for even an educated reader to understand what Wikipedia is talking about. Better to just omit it, IMHO. — Preceding unsigned comment added by Rolf h nelson (talk • contribs) 22:10, 11 January 2014 (UTC)
 * I'm not really qualified to have this argument, as I don't really understand the math behind the calculation. That said,Numberphile did a youtube video on this, in the external links section, which suggests that it is in fact relevant, as it suggests that terminal vacuum may not be the final end state of the universe, and that, given enough time, the universe could reform itself. If that is the case, then it is worth noting.  Serendi <sup style="color:#bb0000;">pod ous  22:44, 11 January 2014 (UTC)
 * Even if that self-published (non-expert?) video carried WP:WEIGHT, it would only support inclusion in an article called "large numbers that have appeared in scholarly papers", since that's what the video is about; the interviewee states he's not endorsing the paper and that the paper hasn't had much influence and jokingly calls one number "a bit of a cheat". Can we find a citation to establish more WP:WEIGHT that the current (IMHO parenthetical) mention in Don Page's paper? If not, I think the material should be removed. Rolf H Nelson (talk) 22:43, 13 January 2014 (UTC)

Date conventions
A portion of this date lists dates however the convention is incorrect. When using "AD" in a date, it should be placed before the year. For example: AD 2014. — Preceding unsigned comment added by Mfigroid (talk • contribs) 00:12, 18 January 2014 (UTC)

Unix time stamps are deterministic, even if they use 64 bits
I have undone this edit for the same reasons that I rejected the exact same logic here:

Redirects for discussion/Log/2013 December 31

Specifically, the claim in the edit summary that
 * "Specifically because this item _relies_ on 'year' being measured using the current time measurement of an earth year"

is incorrect. If you have a counter which overflows at a certain count and a deterministic method of converting that number to years or any other arbitrary units, you can say that it overflows in year X, even in the far future when the earth no longer exists.

If Unix time counted Leap seconds, the argument would be valid, because in that case there would not be a deterministic method of converting Unix time into Coordinated Universal Time, but because the Unix count of since 00:00:00 1 January 1970 (UTC) ignores leap seconds and the UTC representation of time (year, day, hour, minute, second) also ignores leap seconds, the two time numbering systems can be converted even in the far, far future. This conversion has nothing to do with the physical world. It is just arithmetic.

Because of the above, it is valid to calculate the precise second when, say, a 1024-bit seconds counter overflows and put that figure into this article. (We don't actually include that particular figure for another reason; the lack of notable computers that measure time that way).

Could someone who has access to the citation that I restored please post an exact quote of the passage where it discusses 64-but Unix time stamps? I want to confirm the claim that "The source states it can't come true". I suspect that the source does not actually say that. --Guy Macon (talk) 12:00, 26 January 2014 (UTC)
 * Why dont you access the source yourself? It isnt protected in any way.  It is listed at http://www.idrbt.ac.in/workingpapers.html. You shouldnt revert without first reading the source.
 * Anyway, here is the quote:
 * "A 64-bit Unix time would be safe for the indefinite future, as this variable would not overflow until 2**63 or 9,223,372,036,854,775,808 (over nine quintillion) seconds after the beginning of the Unix epoch - corresponding to GMT 15:30:08, Sunday, 4th December, 292,277,026,596.
 * This is a rather artificial and arbitrary date, considering that it is several times the average lifespan of the Sun, the very same celestial body by which we measure time. The Sun is estimated at present to be about four-and-a-half billion years old, and it may last another five billion years before running out of hydrogen and turning into a white-dwarfed star."
 * Years are not pure mathematics. Calculations like this must keep track of the accuracy and precision of the numbers involved.  There is no accuracy problem when talking about storage of a count of time intervals in 64 bits, however there are accuracy problems when introducing the concept of years.  This source does not attempt to predict anything about the far future.  It is stating that 64 bit storage is sufficient for the foreseeable future.
 * But more importantly on this article, unix timestamp is not relevant to this article. It merits a description on the articles about unix timestamp, and may merit listing on time formatting and storage bugs, but it is not in the same ballpark as the other items listed as predictions about events in the far future which are deemed worthy of note.  64bit limits are not worthy of note, based on the list criteria set out in the introduction of this article.  It slipped in during March 2012, and was never discussed.  The near future includes much accelerating change, and many disruptive technologies like quantum computers.  In that context, it is silly to talk of a 64 bit computing in a hundred years from now, let alone millions of years from now - we wont be using computers with a 64bit limit then - there might be one housed in the Interstellar Computer Museum, and they will probably be running a version of linux hacked to have 128-bit time_t (probably Gentoo Linux).  Most modern hardware available now can already do 128-bit storage and computation when required - the decision to use 64bits to hold time was an engineering decision, due to 128bits not being necessary for the foreseeable future. John Vandenberg (chat) 23:59, 26 January 2014 (UTC)


 * Sorry for the confusion about access. I inadvertently clicked on another reference and saw a paywall. Now that I have examined the citation, is is clear that your claim that "the source states it cant come true" misrepresents the source, which says no such thing, and is WP:SYNTHESIS.


 * That being said, your explanation above tells me that we are using two different (yet both valid) definitions of "year". You are talking about a "year" that drifts and gets corrected, I am talking about the "year" that a Unix system displays. Using my definition of year you can calculate the exact second the counter rolls over. With your definition such precision is ridiculous.


 * As far as I can tell, the leap second method only allow a maximum correction of one leap second per day, which is a 0.001% correction, so I changed the page to show less precision than that; "300 billion" instead of "292,277,024,582" and added "approximately" in the text. --Guy Macon (talk) 05:21, 27 January 2014 (UTC)


 * The year that unix systems displays is not time_t - it is struct tm member tm_year, which is a signed int. On all my 64bit unix machines, struct tm.tm_year is 32 bits, so it rolls over to a negative number at 2,147,485,548 (try it at home: $ date --date=2147485547-01-01 && date --date=2147485548-01-01) which is ((2^32)/2)+1900.  That is the time when the unix functions which return a year that has rolled around.  It is long before 64 bit time_t runs out of room to store time intervals since 1970 (~seconds).
 * And, you still havent addressed why unix time should be considered relevant to this list. John Vandenberg (chat) 06:49, 27 January 2014 (UTC)


 * I was explaining the assumptions behind my previous reasoning, comparing them with the assumptions behind your reasoning, and agreeing to accept your assumptions. I see no point in continuing to argue in favor of my assumptions after I have agreed that we will use your assumptions. --Guy Macon (talk) 12:51, 27 January 2014 (UTC)


 * I appreciate that you've shifted from using inappropriate precision to using 300 billion years, however which definition of year is used is not relevant. (Julian year (astronomy) exists to provide precision if needed.) The point is that 'year' doesnt just drift slightly beyond the expected life of the sun - it becomes essentially non-deterministic in your formula, because (a) all relevant scientific theories need to be factored into your accuracy calculations, (b) 'year' in unix is tied to time_t only via struct tm and related functions, which are at the mercy of standard bodies and therefore under constant revision as the physical world changes around us, and (c) the year in unix time will roll over at ~2 billion years, when struct tm fails, not 300 billion years when time_t fails. John Vandenberg (chat) 01:59, 28 January 2014 (UTC)

Didn't the IAU abandon the leap second last year? Also, ease up. I have no idea what triggered this, but it's beginning to look a bit like an edit war.  Serendi <sup style="color:#bb0000;">pod ous  08:18, 27 January 2014 (UTC)


 * It is a bit more complicated than that. http://www.ucolick.org/~sla/leapsecs/onlinebib.html has a good overview of the current standards (note the plural)


 * For the purpose of this page, I was giving John the benefit of the doubt and thus edited the entry so that it is still true whether you assume no leap seconds or one leap second per day.


 * When we talk about events in the far future, we cannot assume that any decision made last year won't be undone next year. The earth is very likely to continue slowing (See Earth's rotation and Tidal acceleration), and it is very unlikely that any standards body will stick to a decision made last year if doing so puts the sun directly overhead at midnight.


 * I am going to address the edit warring issue a bit later. WP:BRD and WP:TALKDONTREVERT are relevant. --Guy Macon (talk) 12:51, 27 January 2014 (UTC)

Removal of sourced content
. with this edit you removed properly-sourced content, and you misrepresented what the source says in this edit summary. Please explain your removal.

Please note that your arguments in the section above do not apply. You were arguing (I am paraphrasing here) that 292,277,024,582 years was far too precise given the fact that, as it says in our article on year,
 * "The word 'year' is also used of periods loosely associated but not strictly identical with either the astronomical or the calendar year".

I agreed with you on that and edited the page accordingly.

I am specifically asking why you removed the following: sentence:
 * "In approximately 300 billion years the Unix time stamp will exceed the largest value that can be held in a signed 64-bit integer.

and why you removed the following reference:

Clearly, a signed 64-bit integer Unix time stamp will overflow in approximately 300 billion years. This article uses "years" to describe (with appropriate precision) events that are much farther in the future than that.

I am asking you to please self-revert your latest edit. --Guy Macon (talk) 14:00, 27 January 2014 (UTC)
 * That paper is not making a scientific prediction about the future, other than 64bit time_t is sufficient for the foreseeable future. It clearly states the number they provide "is artificial and arbitrary" because the assumptions in their applied mathematics are known to be false ("is several times the average lifespan of the Sun the very same celestial body by which we measure time" (the Sun).  i.e. it has an accuracy of zero.  It is relevant as an illustration of the engineering decisions underpinning time_t only.  It is not a pillar of the paper.  Academics who make predictions about events after the lifetime of the Sun, limit themselves to events happening on a scale beyond the Sun. e.g. It is fine to predict that humans will move out of this Solar System, and include them in this article.  However while it is possible to predict dates of lunar eclipses 8 billion years into the future, and asteroids passing the Earth, that doesnt mean they should be included in this article, no matter how unique they might be.  Because any event predicted to occur 8 billion years into the future should be relevant to the universe 8 billion years into the future.  Nobody believes 64bit time_t will be relevant in 8 billion years from now.
 * So, I will not self-revert based on that source alone. The issue is relevance.  I apologise if my edit summary hasnt been clear enough that the issue is that the number is not relevant here.  It is trivia, at best.  If someone else reinstates this entry with this source, it will need to be reviewed at WP:RS/N, and I am quite happy to move this issue through WP:DR in an orderly manner in order to bring in a wider consensus on this entry on this page. John Vandenberg (chat) 15:09, 27 January 2014 (UTC)

That's based on a number of assumptions. Who's to say our starships won't run on Unix, as someone rather pleasingly said on a Reddit thread about this page? Or that we won't adopt "standard years" equivalent to an Earth year, if and when we travel to the stars?  Serendi <sup style="color:#bb0000;">pod ous  19:16, 27 January 2014 (UTC)
 * Not so. Regarding standard Earth years, if we get off this planet soon, we would probably continue using standard Earth years as we knew them in the 21st century, even if it is not our primary way of recording time - it would remain useful for historical comparisons.  However all bets are off for what a 'standard Earth year' is 250 million years from now.  The factors involved in calculating the accuracy of what a year will be are too great to proceed.  That doesnt stop us making predictions about events beyond that point, and using our current value of year to put those events on a timeline.  But it should stop us making predictions where 'standard Earth year' is part of the equation of the prediction (e.g. time_t will run out of seconds because there are x seconds in a standard Earth year).
 * However I am assuming that the some fool wont run a starship that is using an 8 billion year old operating environment (i.e. one which only supports the C99 Unix standard, without any of the (yet to be published) updates to the standard or including non-standard extensions), which already failed to record time properly 6 billion years before hand. (As pointed out in the section above, ~2 billion years from now  has failed us, so the starship using C99 6 billion years after it has failed is more than silly.)  Placing the C99 standard on the timeline of events 300 billion years later is just odd, especially when modern processors and unix compilers already have arithmetic operations on 128 bit registers (gcc has extensions __uint128_t and __int128_t).  If C99 unix is running starships 2 billion years from now, I assume that the standard will be updated (hopefully more than 40 years before the years roll over and everything goes to shit) to use 64 bits for struct tm member 'year', and time_t will be increased at the same time.  i.e. it doesnt belong on a timeline of events predicted to occur in the far future. (it belongs on a timeline of non-events in the far future, collated from non-peer reviewed working papers). John Vandenberg (chat) 01:13, 28 January 2014 (UTC)

Is, as the url suggests, a "working paper"? If so, on its own, it probably carries neither WP:WEIGHT nor is a WP:RS, and therefore should not be included unless a better source is provided; only then should we discuss what the source is or isn't saying. Rolf H Nelson (talk) 04:23, 29 January 2014 (UTC)


 * http://books.google.com/books?id=APJ7QeR_XPkC&pg=PA66
 * Published textbook, probably a WP:RS. However, the closest quote I can find is "This allows for a time representation for 292 billion years": it's not actually saying that running out will or even might happen. Rolf H Nelson (talk) 06:57, 29 January 2014 (UTC)
 * http://vip.cs.utsa.edu/classes/cs3733f2008/recitations/rec02/usp302.pdf
 * Class handout? Probably no WP:WEIGHT. Also doesn't actually say that this event will or might happen. Rolf H Nelson (talk) 06:57, 29 January 2014 (UTC)
 * http://www.merlyn.demon.co.uk/critdate.htm#AD
 * This does say what you want it to say, but it looks self-published, so no WP:WEIGHT. Rolf H Nelson (talk) 06:57, 29 January 2014 (UTC)
 * --Guy Macon (talk) 05:30, 29 January 2014 (UTC)
 * Wait a minute. "...doesn't actually say that this event will or might happen"??? Are you seriously disputing the fact that a 64-bit counter incrementing once per second will exceed the largest value that can be held in a signed 64-bit integer in approximately 300 billion years? Do you also dispute the fact that a one-digit day-of-month variable will exceed it's largest value before the month ends? Please tell me that I am misreading what you wrote (and if I did, I apologize).
 * BTW, See WP:INTERSPERSE. --Guy Macon (talk) 16:52, 29 January 2014 (UTC)
 * I am not disputing the fact that the abstract Unix 64-bit time_t counter will run out in the unlikely event it still exists in any meaningful sense in 300 billion years. However, the textbook provided does not establish any WP:WEIGHT for this fact. The textbook clearly is not predicting this as a future event. Therefore, to use John Vandenberg's terminology, the textbook does not establish that the event is *relevant*, and furthermore does not allay Vandenberg's concern that the event is trivia. To use an analogy, if an editor wanted to include the statement "The 10000th U.S. presidential election is scheduled for AD 41784" but only provides a source saying "presidential elections happen every four years", surely you'd agree that the source given doesn't provide WP:WEIGHT to the proposed statement, even if it's factually true. Rolf H Nelson (talk) 03:37, 30 January 2014 (UTC)

Standards for inclusion
I'm not clear on what criteria editors are using for inclusion in this list. There are a lot of possible events that could happen in the far future; surely the intent isn't to include all of them. Rolf H Nelson (talk) 03:37, 30 January 2014 (UTC)
 * The rate of increase in this article is glacial, so there's not really any problem with inclusion. There's a simple reason why: lots of people are willing to make predictions, very few are willing to put a date on them.  Serendi <sup style="color:#bb0000;">pod ous  08:02, 30 January 2014 (UTC)
 * True, but we can easily derive a couple of standards for inclusion from general Wikipedia principles. First, the Timeline of the future lists pages for present-3000, 3000-10,000, and above 10,000 (this page). We fudge it a bit by going as low as 8,000 in our astronomical events section, but it "at least 10,000 years from now" is a good rule.


 * I would also say that anything that has a Wikipedia article or major coverage in one is clearly notable enough for inclusion here.


 * I would add the following necessary but not sufficient standard for conclusion; The future date must not depend solely on a specific calender system. For example, what makes day 123 in the year 4,567,890 mildly interesting disappears if you use any of the Hindu calendars, whereas the date at which the Unix time stamp will exceed the largest value that can be held in a signed 64-bit integer can be expressed in any calendar. --Guy Macon (talk) 17:05, 30 January 2014 (UTC)


 * It's a bit confusing, but "8000 years" is "years from now" not "AD".  Serendi <sup style="color:#bb0000;">pod ous  18:20, 30 January 2014 (UTC)


 * Oops! Note to self: Next time, smoke crack after editing Wikipedia... --Guy Macon (talk) 18:26, 30 January 2014 (UTC)