> "1990-12-0u" would match the precision you want only if your
> use case was "I know it was one of the first ten months."
I assume you mean "the first nine months".
> Restricting where u can be used is "creating" complexity.
Sorry, I do not agree with that!
> many expression perhaps don't make semantic sense,
I consider that this would be a huge problem.
> From an implementation perspective its less effort to
> accept u anywhere than to have "rules".
It all depends on WHAT you implement. What you wrote is true if what you
want to implement is nothing more than validation. If you, on the other
hand, want to convert / export /reuse / interpret EDTF input in some way,
then implementation is much easier if we know what we are talking about.
In other words: Interpreting unspecified semantics is more useless than