I don't recall the exact discussion. FWIW, I think that both the
exploration and other parts of the protocol have a similar
situation with explicit listing of addresses vs. using indexes:
the former creates a need to be very sure that we have
synchronized lists at both ends in the correct way. IPv6
addresses are large, so there's a tradeoff in packet size
vs. simplicity. At the moment my opinion is that we should
err on the side of simplicity.
In addition, the exploration, reachability, initialization
and update parts have the added complication that if
we ever want to use these components of the protocol
for other purposes, then explicit addresses may be the
only way, given v4 NATs, v4-v6 translation mappings, etc.
may create different views about the addresses on the
two peers.