overlay:
concepts actually provides a generic
definition of Overlay and then uses P2PSIP Overlay for the concept I
think we want.

concepts defines P2PSIP Overlay Name

I like Henning's "search term" and "hashed search term" (I don't know
whether there is a need to say "hashed-ID" it seems redundant.  "search
ID" has a nice property in that it lets us say something like "when
sending a FETCH, encode the search ID at the top of the destination
list" which leaves the statement entirely generic as to whether we're
discussing something that uses a resource-ID or cleartext as routing IDs.

Peer:
 Client:

These seem reasonably well understood, although "client" is dangerous
since, for TCP, there's a client and server side. Maybe we need
something more descriptive such as "peer client", indicating that this
is a client of a peer. This is probably less important since we won't
use that term all that often.


 Peer-ID: (Perhaps, call it Node-ID because clients may also have a peer-ID)

I think this is something where the concepts draft probably should be
changed.  One of the aspects of reload's client protocol that seems to
generate a lot of confusion is that clients can have IDs.  Calling those
Peer-IDs makes little sense.

peer table