here's as far we have got.

>I agree with you in some sense that this would be easier to read.
>However, as soon as you do this, you basically imply that you are
>inserting host routes, which you do not do. How do I distinguish a
>"route" object that is routed, from one that is not? That is very
>confusing. The whole idea about the component is to show parts of a
>route that have different information but are not routed as such.

Clear there is no way to distinguish a route that is routed from
one that is not routed. But It must also be clear that we are
going to leave in a world with route declared and not announced
(that is already the case in the NSFnet database because you
do not destroy the subset at the same time that you create the
superset).

So if you want to clearly distinguish between was is routed and
what is not why not creating an object that can be called "fortools"
for example and with the same philosophy of a route except that
this fortools will never be introduced in the routing mesh.

Because you propose to have a very complicated guardians 

mechanism just for tools.

Gilles



-------- Logged at  Tue May 24 23:58:43 MET DST 1994  ---------