However, the scheme described above is not really perfect.
Since the beginning of the eighties, the numbers
of sites and computer networks have exploded, so that it is virtually
impossible to keep routing information up to date. Also, many
of the central sites are on the Internet now, where routing is
completely different from a UUCP network. Last but not least, site
names on UUCP networks are generally limited to seven, or at most
eight characters, so there is some shortage of names.
Therefore, host names have been organized in a hierarchy of domains,
as described in section 
:
A domain is a collection of sites that are related in some sense -
be it because they form a proper network (e.g. all machines on a
campus, or all hosts on BITNET), because they all belong
to a certain organization (like the U.S. government), or because
they're simply geographically close. Such a domain may itself be part
of a larger domain, which is in turn part of an even larger domain.
This relationship is described as being a subdomainof the
larger domain. To produce a unique address for each machine,
names inside a domain must be unique. Assume the owners of moria
join a small non-profit organization, say Orcs And Thugs Association,
running a network of UUCP sites and paying to other organizations for
access to internation mail services. The network run by them may be
called orcnet, and because they are a some unspecified sort
of organization, they are located below the org top-level
domain. Thus, moria's proper name would be moria.orcnet.org.
This is called its fully qualified domain name, and
uniquely determines the site.
Now, Janet's address would be janet@moria.orcnet.org. However, most sites also understand it when you use domain names in bang paths, like swim.two.birds!moria.orcnet.org!janet. Note that there are still UUCP sites that are not part of any domain. These are considered part of the pseudo-domain .uucp.
The domain name system brings another type of routing to UUCP-based networks. A domain, say foobar.com, may decide to publish only a number of its UUCP sites, and hide the internal network. Now these public sites may be used as gateways to the internal network: any message to an unknown site below the domain .foobar.com will be sent to one of the gateways, which is assumed to forward the message properly.
This method is called domain-based routing. It is primarily useful for large domains, because this allows it to change its internal routing strategy without updating any map entry at the UUCP Mapping Project. New routing information only has to be propagated to the member sites of the domain, which makes it much more flexible. Inside the domain, any routing scheme may be used: through maps distributed inside the domain, a routing scheme based on the subdomain name, or some sort of geographically based routing.