terminological nit

[Steve Deering <deering@parc.xerox.com>]

In the INRIA draft entitled "supporting unidirectional paths in the Internet",
two forms of receiver access are defined: basic access and subnetwork access.
I think those terms could be improved.  The problem I have with "basic access"
is that in English "basic" sometimes implies "most common", but as I
mentioned in an earlier message, I think what is called "basic access" may
become the exception rather than the rule.  The problem I have with
"subnetwork access" is that it implies that there is only one subnetwork
behind the receiver.  I think the real distinction between the two cases
is simply whether the receiver is a router or a (multihomed) host.

Also, those terms only cover the alternative possibilities at the receiving
end of the simplex link; the same alternatives exist at the sending end.
As I see it, we need to consider all combinations of:

	- transmitter is a host
	- transmitter is a router
	- receiver is a host
	- receiver is a router

where I am using the terms "host" and "router" are as defined for IPv6, i.e.,
a router is an IP device that forwards packets not addressed to itself, and
a host is any other IP device.  It is assumed that both hosts and routers
have interfaces to additional links, in order to obtain full-duplex internet
access.

Steve