Hi Anders,
"I seems quite old-school having to manage DHCPv6 scopes just for implementing a feature which is already built into IPv6. Also, since the book was published, it seems that RDNSS have become a wide spread standard in IPv6 networks and has been adopted
by most other platforms out there - except Windows..."
First IPv6 proposal didn't contain RDNSS and people in IETF always argued, that putting DNS info into RA message is a layer violation (information from the application layer is inserted into a message from the network layer). Thus, the "pristine"
solution, according to many people in IETF, is to use RA just for addressing and routing and DHCPv6 for everything else. (You will hear the same arguments if you asked why DHCPv6 doesn't support prefix information for assigned address and default gateway.)
These religious wars are still present and if you watch NANOG or other mailing lists, you will see it every half year. One camp would like to have feature parity, other camp wants "clean" design.
So Microsoft is actually doing things "right" here. Thus, I don't think that Microsoft changes it's mind. However, I absolutely agree with you that from an operation point of view, it is a problem and the situation could be much better.
M.
