Perfect date
Perfect date
DD-MM-YYYY-HH-MM-SS
Makes no sense!
DD-HH-MM-SS-mm-yy for maximum confusion
Yes I was joking, get a random timestamp in this format and you have no idea what it’s referring to.
DD:HH:MM:SS:mm:yy is even better because it could be a MAC address.
2026-05-10T10:06:09.426792Z on all of their tests.
Z at the end denotes UTC.
2026-05-10T10:06:09.426792Z, 2026-05-10 10:06:09.426792Z,2026-05-1010:06:09.426792, and 2026-05-10 all conform to the standard.
After all the self-important blowhards in the committe were satisified that they had put their fingerprint on the ISO8601 document with bullshit like “year-month-week” format support and signed off, they went home.
The rest stayed behind, waited a few minutes to be safe, and then quickly made RFC3339 like a proper standard.
This is what RFC3339 vs ISO8601 feels like.
Let’s not forget that technically you have to play for ISO8601, despite it being nearly useless as a standard because it allows several incompatible formats to coexist.
Fucking wild.
No idea what you based those claims on, but the spec itself (I have the pdf) and Wikipedia’s summary disagree. ISO8601 allows for YYY-MM-DD yes but it allows for a bunch of silly stuff.
en.m.wikipedia.org/wiki/ISO_8601
Both “2025-W24-4” and “2025‐163” are valid representations of today’s date in ISO8601.
(Also the optional timezone makes it utterly useless.)
The omitting of timezones doesn’t matter to a vast majority of the world, since most countries only have one time zone so I don’t see a reason why that is relevant in most use cases.
ISO is a general standard, it’s in the name and the RFC is created for the internet, that is also in the name/description of the RF.
Using 2025-164 can be handy, I actually use the day of the year to check what invoices from previous year are open since those are the invoices that are due 164 days or more.
ISO8601 is much much looser than RFC3339:
Happily!
So, first epoch time. It’s a pretty robust standard, covers many use cases, has few edge cases… but it’s specifically for machine usage, since it’s not human readable and it’s not reversible into the past (pre-1970).
ISO 8601 (depending on the annum), by the text of the documentation, these are all valid dates:
Etc.
RFC 3339 (& RFC 9557, it’s newest modification) is actually a subset of ISO 8601 and is far more prescriptive. For example you must have a timezone designator. You must have a separator between the date and time. You must use a dash between date elements and a colon between time elements. You can easily add standardized subseconds.
This means that RFC 3339 is much easier to parse and use by both machines and humans.
This page (reddit, I know…) has a great summary, and so in the interest of knowledge and attribution I’ll link it: www.reddit.com/r/…/rfc_3339_versus_iso_8601/
This website allows you to more directly compare the two interactively. ijmacd.github.io/rfc3339-iso8601/
We always show up. This meme is factually incorrect ; when you name your files, do you want them in chronological order ? Yes ? Then use ISO.
European dates are better than American dates, but they still produce nonsensical sorting.
/ instead of colons .