I realized that in RIP I goofed even worse than I thought.

In RIP, a computer always has distance zero to itself.
Two computers (hosts, routers, whatever) on the same LAN have distance
one to each other.

Two computers that have to go through k routers to go from one to the
other have distance (hopcount) (k+1).

So, in the network of the CIS 656 project, spring 2004, 
Router R0 has distance zero to itself. 
It has distance one to all (other) computers on the networks I , II ,
III , IV , V , VI . Including Routers R2 , R3 , R4 , R5 , R7 , R9 .
It has distance two to all (other) computers on networks 
VII , VIII , IX , XI , XII (including the routers R6 , R8 , R10).
etc.

---

In RIP, with Poison Reverse:

I said that if router Rx uses the advertisement from router Ry to determine
the route to network Nz (therefore, via router Ry), then, in its router 
advertisement to router Ry, it reports a distance 16 to network Nz.

That is true, but not strong enough.

The correct statement is (see Forouzan, TCP/IP Protocol Suite, p 365,
I hope he is correct, I have not checked this with any RFC):

If router Rx uses information (e.g. a router advertisement) that comes
in over interface Ay to determine the route to network Nz (packets to
Nx therefore leave through interface Ay), then Rx reports, in all
router advertisements it sends out over interface Ay, a distance 16 to
network Nz.

Example:

In the network of the project in CIS 656, router R0 has routing table:

network:   1   2   3   4   5   6   7   8   9  10  11  12  13  14  16
distance:  1   1   1   1   1   1   2   2   2   3   2   2   3   4  
next hop: DD  DD  DD  DD  DD  DD  R2  R4  R5  R5  R7  R9  R9  R9  R9

(network 16 is 0.0.0.0/0 and this is the default route. It has no
distance and is not advertised).

It is understood that for network X R0 uses the advertisement from R5, 
and for network VII it uses the advertisement from R2.

R0 now sends the same advertisement to R4 as well as R5:

network:   1   2   3   4   5   6   7   8   9  10  11  12  13  14
distance:  1   1   1   1  16   1   2  16  16  16   2   2   3   4  

To routers R4 and R5, network V is reported with distance 16, because 
R0 ``knows about'' network V through interface A5, which is the same 
interface it uses to reach R4 and R5.

To routers R4 and R5, network X is reported with distance 16, because 
R0 ``knows about'' X from router R5, through interface A5.

etc.

Since router R0 ``knows about'' network VII through interface A1 
(router R2), the advertisement from R0 to R3 is

network:   1   2   3   4   5   6   7   8   9  10  11  12  13  14
distance:  1   1   1   1   1  16   2   2   2   3   2   2   3   4  

Note that R0 advertises (to R3) distance 2 to network VII, because it
``knows about'' network VII though interface A1.

To R2, R0 advertises distance 16 to network VII. etc.

---

In RIP, with split horizon:

Same situation. If router Rx ``knows about'' a network or a host through
one of its interfaces, then, if the router Rx does RIP with split
horizon, it will not advertise that host or network through the 
interface through which it learned about the host or network.
See Forouzan p 364.

In the situation of the project network, CIS 656 Spring 2004, if router
R0 does split horizon, it will not advertise any of the networks 
V , VIII , IX , X , to the routers R4 and R5.  

---

This may be asked in the final!

In the final, spring 2004, the description as above is the correct one.
Even if the RFC says differently!
(Unless a student shows me differently very soon!)