> > 2008-08-08/yueu
> I have no idea what "yueu" means.
Formally, it is out of today's BNF. But it wouldn't be counter-intuitive
to introduce it there. In analogy with #312 to #314, I was thinking of "u"
as "unspecified" which could be combined with the BNF's definition of
> longYear = "y" "-"? positiveInteger "e" positiveInteger | [...]
Thus, we get "y" (for longYear) "u" (for unspecified) "e" (for exponent)
"u" (for unspecified). The result is "yueu", that is a longYear with
unspecified digits (including those coming after the "e"), which
semantically has some similarities with "unknown".
Limiting the introduction of new "parts" in the EDTF BNF keeps it simple.
Such a notation consists of "known parts" and would therefore be easily
machine-interpreted using a (very slightly) modified version of today's
BNF. It is also logical and easy to memorize. That's what I mean with