"Denenberg, Ray" <[log in to unmask]>:
> Probably the most obvious example is the expanded representation of a year. 8601 says that a year can be more than four digits, but if so the number of digits must be agreed upon in advance by communicating parties -- i.e. by private agreement, outside of the standard, and the means of communicating such agreement is not provided. You may recall, we (that is, this forum) completely rejected that approach as incompatible with the concept of interoperability, and came up with a separate (and interoperable) solution to represent a year requiring more than four digits. In fact I think this is the only case where we rejected the 8601 approach.
As far as I recall, EDTF's 'Y' notation is only specified for years without other date components, whereas years with additional digits in the standard can be used with months, weeks and days. I never understood the restriction in ISO 8601:2004 anyway, because the only thing that would need to be agreed upon is that the extended format is being used. With a maximum of 5 digits per year, it would even suffice to require the leading sign ('+'/'-') to make the format unambiguous (YYDDD was once valid).