[aprssig] Premptive Digipeating

[Robert Bruninga]

> Naming it "preemptive" is somewhat misleading.
> It does not preempt digipeater nodes... later in the path:

We don't want to turn off the rest of the path because that would violate
the sender's intent, and would cripple some very good applications of this
capability.

> SRC>DST,WIDE3-3,NODE2
> That will do WIDE3-3 repeats all over the place,
> except that the NODE2 would pick it up immediately
> even from first transmission.

True.  Nothing in preemptive digipeating is meant to encourage any further
use of generic digipeating beyond 2 generic hops as is currently the
agreed practice.  The W3-3 above is abusive as it has always been.  The
added NODE2 on the end is not, since it is a single "targeted" dupe on the
end of a really big mess.  So a big mess +1 is not the problem.  The
problem is using the 3-3 big mess before it.

> Also explicit chains will do repeats all over the place:
> SRC>DST,NODE2,NODE3,NODE4

Not really.  It can make at most 3 copies, and at least 1.  This is a huge
savings over even WIDE2-2 which can generate as many as from 4 to 16
copies.  In fact, a 3 hop preemtive path above is even better than an
explicit 3 hop path because it always generates 3 hops, whereas the
Premptive will generate on average 1.5 hops.