At the risk of nitpicking, I prefer RFC 3339, which is a profile of ISO 8601. ISO 8601 provides room for things you typically don't want, e.g. dropping time zone offsets.
While I'm typing this I realise that I've never actually read the ISO 8601 spec, like probably most people here, because it's not free. I cannot even be sure it really exists.
Yes, and I've read the drafts of ISO 8601-1:2019 and 8601-2:2019. But can I rely on them? They were published in 2016. What has changed between then and the official release in 2019?
Also, neither RFC3339 nor Wikipedia discuss features like open-ended time intervals. They are in the draft, but did they make the final cut? And in what form? No idea...
The closest I've found to "official" documentation of the 2019 update is a short summary of the changes, hosted by Library of Congress[1]. Which is nice to have, but I sure wish the real standard was public.
> never actually read the ISO 8601 spec ... because it's not free
I was very surprised to discover this. Why are they charging money for the specification of a date format? I would expect standards like these to be published in public domain.
Aside from the fact that you are conflating copyright with price, note that many standards organizations charge money for copies of their standards documents.
ISO charges roughly 20 times the price of a paperback book for a 33 page document. You can buy BS/ISO 8601-1:2019 from the BSI for an even more exorbitant £246. Standards Australia will sell the older 2007 version to you for a mere AUD165.
But this is not special. All standards documents cost money from these organizations, from dates and times to electrical plugs. Treasure the fact that you can (for example) get (some) ECMA standards from ECMA for free. It isn't the norm.
ISO finances their operations by selling documentation, subscription from members, etc. The point of ISO standards is to facilitate international trade. The intended audience for these standards are large multinational companies.
Pretty much all the standards bodies I know of are financed through sales of the standards documents. It's not a bad system, in the end the commercial entities who end up needing them most pay for their development and upkeep.
I have read the actual ISO 8601 spec. It is somewhat typical wall of prose ISO standard that contains meaningful conformance distinctions which depend on two words in the middle of 10+ line paragraph.
For me, more significant reason to prefer RFC 3339 constrained profile is that ISO 8601 does not only mean YYYY-MM-DDTHH:MM:SS.sss but also includes bunch of alternate syntaxes like YYYY-DDD. And I have one big issue with RFC3339: there are cases when the "don't care about timezone" ISO 8601 syntax is actually useful.
> While I'm typing this I realise that I've never actually read the ISO 8601 spec, like probably most people here, because it's not free. I cannot even be sure it really exists.
Always wondered about this too. How can people implement a standard if they can't freely access and reference it? It's the 21st century. Why can't they just publish the document on the internet just like the IETF does?
> How can people implement a standard if they can't freely access and reference it?
They pay for it, out of the money they expect to make on the implementation.
It's a little problematic for free implementations, but then for most standards much less so than the engineering going into the implementation which has an (opportunity, if nothing else) cost.
While I'm typing this I realise that I've never actually read the ISO 8601 spec, like probably most people here, because it's not free. I cannot even be sure it really exists.