...
Some of that
I still believe could be cleaned up to be nicer. It doesn't
relate to any of the substance though, nor particularly to
BIND 9, but rather to use of terms like "slave" and "master"
which are (recent) BIND inventions to refer to what the DNS
(1034/1035) has always used the words "secondary" and
"primary" for.
These terms were defined in RFC1996 and are also used in
RFC2136, and they seem clearer than "primary" and "secondary"
when describing topologies where a slave server is itself
acting as a master for other slaves. If you like, I could add
a reference like "This document uses the terms master and
slave as defined in RFC1996".
I'm not interested in, and probably not competent to, engage in
the larger argument here, but I think that definitional
clarifications like that one, which explain what the terms you
are using mean, are _always_ desirable.