02:10 <jfhbrook> a sol is obviously a year, not a day

02:10 <jfhbrook> >:O

03:14 <ELLIOTTCABLE> jfhbrook: wut

03:14 <ELLIOTTCABLE> the point of the system is that, out of all the possible reference-points for time, since we have to choose a *single one*,

03:14 <ELLIOTTCABLE> (i.e. if we're gonna use a metric system, we have to have *something* that we're multiplying by ten, 100, 0.0001, whatever)

03:15 <ELLIOTTCABLE> then the most *necessary* one (i.e. the one we can't work around, change, or live without) is that of our circadian rhythm - our sleep pattern

03:15 <ELLIOTTCABLE> as someone who's on-and-off tried polyphasic sleep, and dealt with insomnia his *entire life*, trust me, there's absolutely zero chance of us getting rid of the "solar day" within the next 10,000 years.

03:16 <ELLIOTTCABLE> whereas the solar *year* is easily disposed of - we have A/C and central heating; honestly, some specific professions (farming, meterology) aside, the solar year has zero effect on the average human's life. (at least when compared to the effect that the solar day has.)

03:16 <ELLIOTTCABLE> meanwhile the second is laughably fucking pointless in every way, can we plz get rid of that A.S.A.P.

03:17 <ELLIOTTCABLE> so given that the only reasonable basis for a metric timekeeping system is the human-sleep "day" (we could try and be pedantic and match it to the Average of All Living Humans' Circadian Rhythyms, which I understand to be something like 23 hours and 10 minutes? but meh, more useful to ease transition by matching it precisely to 24 hours),

03:17 <ELLIOTTCABLE> the only remaining choice to make is basically the epoch.

03:17 <ELLIOTTCABLE> and I think UNIX chose that pretty well: all things of *real importance* started when computers did.

03:17 <ELLIOTTCABLE> controversial statement, but smehidontcare.

03:17 <ELLIOTTCABLE> so: janurary 1st 1970, plus the 24-hour solar day, and boom you have metric time.

03:50 <ljharb> ELLIOTTCABLE: a metric of time needs to be objectively measureable tho. you can easily measure a solar year. you can't measure "average human sleep length"

03:51 <ELLIOTTCABLE> not necessarily at all

03:51 <ljharb> you can measure a solar or lunar anything, obv, and "solar day" and "solar year" and "lunar month" are the easiest

03:51 <ELLIOTTCABLE> first off solar year changes just as much as human sleep length, on the grand scale of things

03:51 <ELLIOTTCABLE> waaaaaaay too much to be a *permanent* basis

03:51 <ELLIOTTCABLE> whatever the choice, it's gonna have to be redefined in vibrations of a cesium atom or w/e

03:51 <ljharb> it does change, but on a relative scale, human sleep changes more.

03:51 <ljharb> ie the percentage of a year that the solar year deviates by is small (even if it adds up over time)

03:51 <ELLIOTTCABLE> and might as well measure 10,000 humans and take the average, as measure 10,000 years and take the average

03:52 <ljharb> it's easier to measure 10,000 years, and have multiple independent observers corroborate.

03:52 <ELLIOTTCABLE> and the 10,000 humans are 10,000 times more relevant to, well, humans :P

03:52 <ljharb> measuring anything by "humans" is just a flawed idea.

03:52 <ELLIOTTCABLE> so? doesn't matter. still less relevant.

03:52 <ljharb> it's relevant, but it's not objectively measureable

03:52 <ELLIOTTCABLE> no matter what, it's not *stable enough* to have a *meaningful* level of stability,

03:52 <ljharb> and objective measurement is the most relevant.

03:52 <ljharb> it's been stable enough for us for a long time

03:52 <ELLIOTTCABLE> and thus we might as well throw stability out entirely and measure something actually fucking useful

03:52 <ljharb> human sleep isn't stable for even 1 person, over a long enough period of time.

03:53 <ELLIOTTCABLE> the only relevant amount of stability is "this will be the same length of time when, in 10,000 years, we're measuring things against this, we're talking about almost the same period of time"

03:53 <ljharb> don't get me wrong, i'm not saying that if we reinvented "measuring time" now, we'd end up with anything like what we've got

03:53 <ljharb> right, and that's false

03:53 <ELLIOTTCABLE> solar cycle ain't remotely stable enough for that - cesium atom is

03:53 <ljharb> after 10,000 years of exposure to electric lights, human sleep will be wildly different

03:53 <ljharb> now, cesium atoms, true

03:53 <ELLIOTTCABLE> disagree strongly

03:54 <ELLIOTTCABLE> we've been exposed to all sorts of changes and it's *barely* changed

03:54 <ljharb> and absent historical baggage, measuring time in terms of objectively measureable physical laws like atomic decay is the right approach

03:54 <ljharb> um, it's changed a *lot* in the last 400 years

03:54 <ELLIOTTCABLE> our sleep-cycle is barely distinguishable from that of Neolithic humans; and probably the Single Largest Area of space-living research, besides pooping, is how to make *space living* work around *our* sleep schedules

03:54 <ljharb> prior to that, humans slept 4-6 hours from soon after sunset, woke up for 2-3 hours, and then slept til sunrise

03:54 <ljharb> our sleep cycle is highly distinguishable from every human on the planet before the 1400s.

03:54 <ELLIOTTCABLE> because they're damn-straight not gonna even *begin* to change before we've got like, fifty human colonies on distant planets with wormhole travel.

03:54 <ELLIOTTCABLE> they just. don't. change. ever.

03:55 <ELLIOTTCABLE> you and I have read *wildly* differing sources, then :P

03:55 <ljharb> i'm having a baby in a few months and i just finished multiple long long books on sleep

03:55 <ELLIOTTCABLE> anyway *shrug* I cared about this in, like 2009

03:55 <ELLIOTTCABLE> I enjoy UJD, but it's literally a channel joke at this point :P

03:55 <ljharb> ujd?

03:55 <ELLIOTTCABLE> not going anywhere, so not remotely worth drunkenly arguing :P

03:55 <ljharb> oh that link

03:55 <ELLIOTTCABLE> the time-system I described above

03:55 <ljharb> reading

03:56 <ELLIOTTCABLE> before wanting to Reinvent Asynchronous Programming with paws, I wanted to Reinvent Timekeeping with ujd

03:56 <ELLIOTTCABLE> :P

03:56 <ljharb> ok so your time system here

03:56 <ELLIOTTCABLE> there's nothing at the link

03:56 <ELLIOTTCABLE> it's literally just what I said above:

03:56 <ELLIOTTCABLE> Julian date, on the UNIX epoch

03:56 <ljharb> so 0 is unix 0?

03:56 <ELLIOTTCABLE> slightly re-defined on the cesium atom instead of the actual sun

03:56 <ljharb> so what does `1` in the rightmost group represent?

03:57 <ELLIOTTCABLE> yah, it's literally unix_timestamp / 86400) right now

03:57 <ELLIOTTCABLE> there'll need to be some leap-second adjustments in a few years, but *shrug*

03:57 <ELLIOTTCABLE> another upside, thrown in with the dozen I've already got: easy to calculate on existing systems!

03:57 <ELLIOTTCABLE> lmao

03:57 <ljharb> ok so the rightmost group is "increments every time a cesium atom lets loose a particle" or something?

03:58 <ljharb> and each other group is just 1000 of the group to the left?

04:34 <ELLIOTTCABLE> group? what?

04:35 <ELLIOTTCABLE> 17 102 191 058 --> 17,102.191058 sol since epoch (basically days)

04:37 <ELLIOTTCABLE> 1 sol == "the duration of 794243384928000 periods of the radiation corresponding to the transition between the two hyperfine levels of the ground state of the caesium 133 atom", for instance

04:38 <ljharb> ohhh

04:38 <ljharb> the spaces implied separation like hours/minutes/seconds

04:38 <ljharb> also the inline units, like the ƒ

05:57 <ELLIOTTCABLE> ljharb: oh, but there are!

05:57 <ljharb> lol ok that's what i'm trying to understand

05:57 <ELLIOTTCABLE> ahhah i'm wildly depressed and I just sat down at the computer, opened Spotify

05:57 <ELLIOTTCABLE> and it starts playing Hello - Evanescence

05:57 <ELLIOTTCABLE> U NO ME 2 WELL, SPERTERFER

05:58 <ELLIOTTCABLE> 22222 WELL