I'm going to be contrarian.
I think ietf(_at_)ietf(_dot_)org is *exactly* what we want for IETF last
calls. Most last calls are silent. Some trigger a small,
non-annoying amount of technical discussion. The remaining
ones cause mailstorms. Those are exactly the ones I, as an
IETF citizen, want to know about. They tell me that the IETF
is about to do something controversial, and I need to have
a careful look to see if I care. If I decide that I don't care,
it's trivial to ignore the thread.
This essential feature would be lost if the last call traffic
was hidden in some place dedicated to the particular draft;
I'd never be aware that there was a controversy.
To say that another way: a last call message on IETF-announce
would at most attract the attention of people who already care.
A last call mailstorm here will attract the attention of
people who ought to care, and slightly interrupt the viewing
experience of people who don't care.
That said, I'm all for attempting to dissuade inappropriate
messages during such a mailstorm. But the mailstorm itself
has value.
Brian