ISO is problematic

Posted Apr 13, 2023 16:54 UTC (Thu) by flussence (guest, #85566)
In reply to: ISO is problematic by chris_se
Parent article: Standardizing BPF

One time I went to actually look up ISO 8601 out of curiosity, and that's how I found out everyone on the internet was cargo-culting it. The more interesting spec is RFC 3339, a subset of ISO 8601 that contains just the date formatting part, and none of the sanity-eating nonsense like calendar-relative duration shorthand. (It would be nice if ostensibly-open formats like WHATWG HTML would reference this one instead of a loginwall…)

In any case, I don't think that organisation has been fit to define standards since the advent of ISO 29500 (OOXML).

to post comments

ISO is problematic

Posted Apr 13, 2023 18:12 UTC (Thu) by excors (subscriber, #95769) [Link]

> It would be nice if ostensibly-open formats like WHATWG HTML would reference this one instead of a loginwall…

HTML only mentions ISO 8601 as an explicitly non-normative reference; you don't need it in order to use or implement HTML, it's just referenced as background information for some of the features normatively defined within the HTML spec. In one case it's referenced specifically to warn against using an ISO 8601 implementation, because ISO doesn't define the parsing rules properly and you need to implement HTML's parsing rules instead.

HTML supports week strings (<time datetime=2023-W15>this week</time>) which aren't defined in RFC 3339 so that wouldn't be a useful reference.

ISO is problematic

Posted Apr 13, 2023 18:30 UTC (Thu) by kunitz (subscriber, #3965) [Link]

Be careful with RFC 3339. The RFC does specify datetime stamps that are valid ISO 8601 datetime stamps, but are only a subset. Be also aware that the ABNF in the appendix of the RFC describes ISO9601:1998 time stamps, which differs from the ABNF actually defined in the text of the RFC. ISO 8601 allows two-digit years and defines the comma as preferred decimal sign, but allows the dot too.