> This is true of RIPv1, but not RIPv2 which handles variable
> length masks.  Our routed(8) supports both.

Ah, yes, but folks sticking with RIP probably mostly do so
because they have legacy devices hanging around which don't do
RIPv2, so they're basically stuck with RIPv1.  I sure know that
if I was given a choice between OSPF and RIPv2 I would not
hesitate to choose OSPF.

> For a free implementation of OSPF you could always port the
> kernelized version presented by the 'Xinu' operating system (as
> described by 'Internetworking with TCP/IP' volume 2).
>
> 	ftp://ftp.cs.purdue.edu/pub/Xinu

Interesting.

- H=E5vard