I just use millis since epoch
(Recently learned that this isn’t accurate because it disguises leap seconds. The standard was fucked from the start)
This is the way.
How is day smaller than month? There are up to 12 values for month, but up to 31 for days
obvious troll is obvious
It’s sorted by the length of time, so a day is shorter than a month.
I use ss/mm/hh/dd/MM/YYYY
t.european
Hot take: 2025-Jan-27 is better than 2025-01-27 in monolingual contexts.
The beautiful part of 2025/01/27 is that it can inherently be sorted without formatting.
finally a correct version of this diagram
i never saw year first in Europe.
You’re reading the post backwards.
Mmm US military date and time is fun too.
DDMMMYYYYHHMM and time zone identifier. So 26JAN20251841Z.
So much fun.
So virtually human unreadable and the letters make machine readability a pain in the ass?
As my friend used to say, there’s dumb and then there’s Army Dumb.
Honestly look very readable to me, though I’m not sure on the timezone bit. Maybe they left it out? Ohterwise it’s 26th of January 2025, 18:41
It’s gonna be problematic when there’s 5 digit years, but other than that it’s… not good, but definitely less ambiguous than any “normally formatted” date where DD <= 12. Is it MM/DD or DD/MM? We’ll never fucking know!
Of course, YYYY-MM-DD is still the king because it’s both human readable and sortable as a regular string without converting it into a datetime object or anything.
All you’d have to do to make it much more readable is separate the time and the year with some kind of separator like a hyphen, slash or dot. Also “Z” is the time zone, denoting UTC (see also military time zones)
Oh, duh. It’s why all my timestamps have Z’s in the database lmao
Thing is, you’re right that the separation would help, but this is still way less ambiguous that MM/DD vs DD/MM if you ask me.
All my homies hate ISO, RFC 3339 for the win.
All my homies hate ISO
Said no-one ever?
Nah, ISO is a shit organization. The biggest issue is that all of their “standards” are blocked behind paywalls and can’t be shared. This creates problems for open source projects that want to implement it because it inherently limits how many people are actually able to look at the standard. Compare to RFC, which always has been free. And not only that, it also has most of the standards that the internet is built upon (like HTTP and TCP, just to name a few).
Besides that, they happily looked away when members were openly taking bribes from Microsoft during the standardization of OOXML.
In any case, ISO-8601 is a garbage standard.
P1Yis a valid ISO-8601 string. Good luck figuring out what that means. Here’s a more comprehensive page demonstrating just how stupid ISO-8601 is: https://github.com/IJMacD/rfc3339-iso8601P1Y is period notation. It means a Period of 1 Year. It actually makes decent sense tbh.
Sure, it means something, and the meaning is not stupid. But since it is the same standard, it should be possible to be used to at least somehow represent the same data. Which it doesn’t.
I think it is reasonable to say: “for all representation of times (points in time, intervals and sets of points or intervals etc) we follow the same standard”.
The alternative would be using one standard for points in time, another for intervals, another for time differences, another for changes to a timezone, another for …
The alternative would be
More reasonable, if you ask me. At least I came to value modularity in programming, maybe with standards it doesn’t work as good, but I don’t see why
Standards are used to increase interoperability between systems. The more different standards a single system needs the harder it is to interface with other systems. If you have to define a list of 50 standard you use, chances are the other system uses a different standard for at least one of them. Much easier if you rely on only a handful instead
if i am not wrong, it is because essentially both are same (slight differences in what is allowed and what is not, https://github.com/IJMacD/rfc3339-iso8601), but RFC is more free as in freedom
Thx i take that back
I work with international clients and use 2025-01-26 format. Without it… confusion.
That’s an ISO date, and it’s gorgeous. It’s the only way I’ll accept working with dates and timezones, though I’ll make am exception for end-user facing output, and format it according to locale if I’m positive they’re not going to feed into some other app.
This is Belize and Micronesia erasure.
I don’t know why anyone would ever argue against this. Least precise to most precise. Like every other number we use.
(I don’t know if this is true for EVERY numerical measure, but I’m sure someone will let me know of one that doesn’t)
They are all equally prescise. American one is stupid just like their stupid ass imperial units. European one is two systems slapped together(since they are rarely used together and when they are its the iso format) and iso is what european standard should be.
You misunderstand my comment.
I’m saying the digits in a date should be printed in an order dictated by which units give the most precision.
A year is the least precise, a month is the next least, followed by day, hour, minute, second, millisecond.
You are looking not for precision but for largest to smallest, descending order. this is distinct from precision, a measure of how finely measured something is. 2025.07397 is actually more precise than 2025/01/27, but is measured by the largest increment.
And to address the argument on precision versus descending. I disagree. An instrument counting seconds is more precise than a machine counting minutes, hours, days, weeks, months etc… And that holds true through the chain. The precision is in the unit.
We can debate this all day. And I can’t honestly say that I would take either side in a purely semantics argument.
But the wording comes directly from RFC3339 which is, to me, the definitive source for useful date representation.
https://www.ietf.org/rfc/rfc3339.txt
5.1. Ordering
If date and time components are ordered from least precise to most precise, then a useful property is achieved. Assuming that the time zones of the dates and times are the same (e.g., all in UTC), expressed using the same string (e.g., all “Z” or all “+00:00”), and all times have the same number of fractional second digits, then the date and time strings may be sorted as strings (e.g., using the strcmp() function in C) and a time-ordered sequence will result.
Largest to smallest is also wrong. In 2025/01/28, the 28 is larger than the 01.
It should be “most significant” to “least significant”
largest to smallest is correct. 1 mile is larger than 20 meters. if i had specified numerical value or somesuch, maybe you’d be correct. though significance works as well.
Largest to smallest is at best ambiguous. It can refer to the size of the number itself, or the size of the unit.
There is a reason this exact concept in maths/computer science is known as the “significance” of the digit. Eg. The “least significant bit” in binary is the last one.
Sorting with either the month or the day ahead of the year results in more immediately relevant identifiable information being displayed first. The year doesn’t change very often, so it’s not something you necessarily need to scan past for every entry. The hour changes so frequently as to be irrelevant in many cases. Both the month and the day represent a more useful range of time that you might want to see immediately.
Personally, I find the month first to be more practical because it tells you how relatively recent something is on a scale that actually lasts a while. Going day first means if you’ve got files sorted this way you’re going to have days of the month listed more prominently than months themselves, so the first of January through the first of December will all be closer together then the first and second of January in your list. Impractical.
Year first makes sense if you’re keeping a list around for multiple years, but the application there is less useful in the short term. It’s probably simpler to just have individual folders for years and then also tack it on after days to make sure it’s not missing.
Also, like, this format is how physical calendars work assuming you don’t have a whole stack of them sitting in front of you.
By keeping years in different folders you are just implicitly creating the ISO format: eg.
2025/"04/28.xls"
Why would the year, the least important, need to be first?
And why are the pieces of the pyramid made so the ISO standard is the only one that looks right? ss:mm:hh:DD:MM:YYYY would also order the numbers based on length, but would look terrible if represented like that
Why would the year, the least important, need to be first?
Maybe it’s not least important for everyone?
Almost like the preference can change depending on application…
If I’m looking at a folder full of spreadsheets, one each month (or even day) for several years, and they are all titled according to YYYY-MM-DD. All you need to do is sort by filename and now you have it broken down by year, into one spreadsheet per month/day.
And only needed to click one button to sort them into an easily readable format.
I often see the year being the most important in my archive. Followed by month, then day (which is often left out because the document is monthly).
And the why; because it sorts alphabetically.
Well you read the least frequently changing part first, the year, because if you read the seconds first, then the thing’s already changed before you’ve even finished reading it. /s
Why would the year, the least important, need to be first?
For proper ordering for one. ISO8601 is objectively the best way to label anything that might need to be ordered based on time. This forces data points to line up properly in chronological order, and makes it easy to time slice as needed.
And why are the pieces of the pyramid made so the ISO standard is the only one that looks right?
Because it’s the only one that goes from largest value to smallest. It’s first because you start from the largest as the base (year) and work down through size to seconds.
ss:mm:hh:DD:MM:YYYY would also order the numbers based on length, but would look terrible if represented like that
Agreed. And any sort of data analysis would be so much harder
Arent there uses other than ordering files?
The ISO standard is best for ordering files, but that doesnt mean its good for other things
Its impossible to confuse it with the other 2 presented in this post so you could use it for files and use another one for other things
Edit: i may have been misunderstanding the context in which the ISO standard is claimed to be superior
The fact that you can’t confuse it with other formats is precisely the advantage. With any other format (besides the awful lettered month) you have to use context clues to be sure you’re reading it correctly if the day is less than 13.
Europe: 10/12/2025 USA: 12/10/2025 If you don’t have context as to which system this is, would 2025/12/10 make things less ambiguous?
To be fair, proper ISO 8601 specifies hyphens as the separator between date elements, and I don’t think I’ve ever seen a XXXX-XX-XX (with hyphens) be used for YYYY-DD-MM. Just XX-XX could perhaps be ambiguous, but fortunately that’s not allowed by the standard, and anyone using just year-day for XXXX-XX is absolutely trolling. YYYY-DDD could have a use, though should really use a separate separator to not sort together IMO. A year-week designation could possibly look like XXXX-XX, but that seems unlikely to just be dropped in that format without context, at least to my western US sensibilities.
You can skip the year and just do 1-26
Sometimes you need the year
Yeah it’s funny seeing everyone in here thinking only of the one specific thing they use this for, not recognizing that the most useful order can change depending on the purpose of the data.
I know, why don’t we all agree to agree and use every single possible format within a shared spreadsheet
Oh god it’s so true…
My stupid ass read this top to bottom and I was confused why anyone would start with seconds
