Probably we are following the Wikipedia standard, since most if not all the libraries that everyone uses to parse and produce ISO dates were written by people t

Probably we are following the Wikipedia standard, since most if not all the libraries that everyone uses to parse and produce ISO dates were written by people that read Wikipedia and not by people that bought the standard. The problem with "closed" standard is that most people if they have to pay for something they find other way, and assume the standard by reading other sources, Wikipedia, the implementation of some library that is supposed to implement the standard, or the drafts of the standard (like it's done with the C standard, do you assume that every developer of gcc have bought the standard?). A standard isn't a standard because someone wrote a paper and sold it, a standard is a standard if it's used by people. And it happens that since standards are not freely available people makes things that are close but not really conform to the standard, and these things because the de-facto standard. And you developer you have to support these cases. Let's say ISO8601, there are a ton of applications that produces dates that are similar to ISO8601 format but not quite like it. And that you must support because they are so common... and then you find piece of codes that tries to interpret a date in various ways trying to compensate for minor standard errors. It's a mess that could have been avoided if the standard document were really open.