> The ability of these roots to live together has been explained by Karl
> Auerbach:
>
> 2.  Multiple Roots are "a good thing"
> http://www.cavebear.com/cavebear/growl/issue_2.htm#multiple_roots

But Karl is wrong. He is comparing the DNS with phonebook publishers.
The difference is that the phonebook publishers are not 'authoritative'
over the data (the phonenumbers) in their phonebooks.
The assignment of telephone numbers is controlled by the National
Numbering Plan Administrators in each country. And there is also only
one, and their common root is the ITU which allocated country codes.
If each telco was free to assign their own area codes, telephone numbers
etc that would be much more akin to multiple roots than phonebook
publishers.

The publishing of phonebooks is more akin to search engines and index pages,
that simply list (arbitrary subsets) of existing allocations, but have to
power over the existence, allocation or functioning of same.

He also says:
> From a technical point of view all that a root server group does is to
give
> its users a way to find the DNS servers that handle the various Top Level
Domains
> (TLDs).  The root servers do not themselves answer queries about what
names are
> inside the various TLDs.  Those questions are passed on to the TLD servers
themselves.

> That is a subtle point and a point that is often lost when discussing the
DNS.

> It bears repeating -- all that a root server does is to answer queries
about how to
> find a server handling a TLD named in the query.  In other words, a root
server only
> answers queries such as "Where do I find a server that contains the list
of names in .com?".

But that is exactly the problem. How can it be ensured that the 'alternate'
roots point
to the correct TLD nameservers. He seems to suggest that by simply running a
root-ns
oneself, it magically would do the right thing. NOT SO. The coordination
between these
is the problem in question and no amount of handwaving will make that go
away.

Mathias